United States Solid Waste And
Environmental Protection Emergency Response/
Agency Research And Development
(OS-420) WF
EPA/530/UST-90/010
September 1990
Standard Test Procedures
For Evaluating Leak
Detection Methods
Pipeline Leak Detection
Systems
Printed on Recycled Paper
-------
Standard Test Procedures for
Evaluating Leak Detection Methods:
Pipeline Leak Detection Systems
Final Report
U.S. Environmental Protection Agency
Office of Research and Development
September 1990
-------
DISCLAIMER
This material has been funded wholly or in part by the United States Environmental
Protection Agency under contract 68-03-3409 to COM Federal Programs Corporation. It
has been subject to the Agency's review and it has been approved for publication as an
EPA document. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial products frequently
carry with them the increased generation of materials that, if improperly dealt with, can
threaten both public health and the environment. The U.S. Environmental Protection
Agency is charged by Congress with protecting the Nation's land, air, and water
resources. Under a mandate of national environmental laws, the agency strives to
formulate and implement actions leading to a compatible balance between human
activities and the ability of natural systems to support and nurture life. These laws direct
the EPA to perform research to define our environmental problems, measure the
impacts, and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for planning, implementing,
and managing research, development, and demonstration programs to provide an
authoritative, defensible engineering basis in support of the policies, programs, and
regulations of the EPA with respect to drinking water, wastewater, pesticides, toxic
substances, solid and hazardous wastes, and Superfund-related activities. This
publication is one of the products of that research and provides a vital communication
link between the researcher and the user community.
Risk Reduction Engineering Laboratory
E. Timothy Oppelt, Director
-------
PREFACE
Part of a series published by the Environmental Protection Agency (EPA) on standard
test procedures for evaluating leak detection methods, this document addresses how to
evaluate leak detection systems designed for pipelines associated with underground
storage tanks.
How to Demonstrate That Leak Detection Methods Meet EPA's Performance
Standards
The EPA's regulations for underground storage tanks require owners and operators to
check for leaks on a routine basis using one of a number of detection methods (40 CFR
Part 280, Subpart D). In order to ensure the effectiveness of these methods, the EPA
has set minimum performance standards for equipment used to comply with the
regulations. For example, after December 22, 1990, all systems that are used to
perform a tightness test on a tank or a pipeline must be capable of detecting a leak as
small as 0.10 gallons per hour with a probability of detection of at least 95% and a
probability of false alarm of no more than 5%. It is up to tank owners and operators to
select a method of leak detection that has been shown to meet the relevant performance
standard.
Deciding whether a system meets the standards has not been easy. Until recently,
manufacturers of leak detection systems have tested their equipment using a wide
variety of approaches, some more rigorous than others. Tank owners and operators
have been generally unable to sort through the conflicting sales claims based on the
results of these evaluations. To help protect consumers, some state agencies have
developed mechanisms for approving leak detection systems. These approval
procedures vary from state to state, making it difficult for manufacturers to conclusively
prove the effectiveness of their systems nationwide. The purpose of this document is to
describe the ways that tank owners and operators can check that the leak detection
equipment or service they purchase meets the federal regulatory requirements. States
may have additional requirements.
The EPA will not test, certify, or approve specific brands of commercial leak detection
equipment. The large number of commercially available leak detection systems and
methods makes it impossible for the Agency to test all the equipment or to review all the
performance claims. Instead, the Agency has described how equipment should be
tested to prove that it meets the standards. This testing process is called the evaluation,
the results of which are summarized in a report. The information in this report is
intended to be provided to customers or regulators upon request. Tank owners and
operators should keep the evaluation results on file to satisfy the EPA's record-keeping
requirements.
-------
The EPA recognizes three distinct ways to prove that a particular brand of leak detection
equipment meets the federal performance standards:
1. Evaluate the method using the EPA's test procedures for leak detection
equipment.
2. Evaluate the method using a voluntary consensus code or standard
developed by a nationally recognized association or independent third-party
testing laboratory.
3. Evaluate the method using a procedure deemed equivalent to the EPA
procedure by a nationally recognized association or independent third-party
testing laboratory.
Manufacturers should use one of these three approaches to prove that their systems
meet the regulatory performance standards. For regulatory enforcement purposes, each
of the approaches is equally satisfactory.
EPA Test Procedures
The EPA has developed a series of test procedures that cover most of the methods
commonly used for underground storage tank leak detection. The particular procedures
for each type of system or method are described in a report that is part of a larger series.
The series includes:
1. "Standard Test Procedures for Evaluating Leak Detection Methods:
Volumetric Tank Tightness Testing Methods"
2. "Standard Test Procedures for Evaluating Leak Detection Methods:
Nonvolumetric Tank Tightness Testing Methods"
3. "Standard Test Procedures for Evaluating Leak Detection Methods:
Automatic Tank Gauging Systems"
4. "Standard Test Procedures for Evaluating Leak Detection Methods:
Statistical Inventory Reconciliation Methods"
5. "Standard Test Procedures for Evaluating Leak Detection Methods: Vapor-
phase Out-of-tank Product Detectors"
6. "Standard Test Procedures for Evaluating Leak Detection Methods: Liquid-
phase Out-of-tank Product Detectors"
7. "Standard Test Procedures for Evaluating Leak Detection Methods: Pipeline
Leak Detection Systems"
IV
-------
Each report on a type of system or method provides an explanation of how to conduct
the test, how to perform the required calculations, and how to report the results. The
results from each standard test procedure provide the information needed by tank
owners and operators to determine whether the method meets the regulatory
requirements.
The EPA test procedures may be used either directly by equipment manufacturers or by
an independent third party under contract to the manufacturer. Both state agencies and
tank owners typically prefer a third-party evaluation, since this is a more objective way of
demonstrating compliance with the regulations. Independent third parties may include
consulting firms, test laboratories, not-for-profit research organizations, or educational
institutions with no organizational conflict of interest. In general, the EPA believes that
the greater the independence of the evaluating organization, the more likely it is that an
evaluation will be fair and objective.
National Consensus Code or Standard
A second way for a manufacturer to prove the performance of leak detection equipment
is to evaluate the system according to a voluntary consensus code or standard
developed by a nationally recognized association (American Society of Testing Materials
(ASTM), American Society of Mechanical Engineers (ASME), American National
Standards Institute (ANSI), etc.). Throughout the technical regulations for underground
storage tanks, the EPA has relied on national voluntary consensus codes to help tank
owners decide which brands of equipment are acceptable. Although no such code
presently exists for evaluating leak detection equipment, one is under consideration by
the ASTM D-34 subcommittee. The Agency will accept the results of evaluations
conducted according to this or similar codes as soon as they have been adopted.
Guidelines for developing these standards may be found in the U.S. Department of
Commerce's "Procedures for the Development of Voluntary Product Standards" (FR,
Vol. 51, No. 118, June 20, 1986) and OMB Circular No. A-119.
Alternative Test Procedures Deemed Equivalent to the EPA's
In some cases, a leak detection system may not be adequately covered by EPA
standard test procedures or a national voluntary consensus code, or the manufacturer
may have access to data that make it easier to evaluate the system another way.
Manufacturers who wish to have their equipment tested according to a different plan (or
who have already done so) must have that plan developed or reviewed by a nationally
recognized association or independent third-party testing laboratory (Factory Mutual,
National Sanitation Foundation, Underwriters Laboratory, etc.). The results should
include a certification by the association or laboratory that the conditions under which the
test was conducted were at least as rigorous as the EPA standard test procedure. In
general this will require the following:
-------
1. The system should be tested on an underground storage tank or associated
pipeline both under the no-leak condition and an induced-leak condition with
an induced leak rate as close as possible to (or smaller than) the EPA
performance standard. In the case of tank or pipeline tightness testing, for
example, this will mean testing under both 0.0-gal/h and 0.10-gal/h leak rates.
In the case of groundwater monitoring, this will mean testing with 0.0 and
0.125 in. of free product.
2. The system should be tested under at least as many different environmental
conditions as are included in the corresponding EPA test procedure.
3. The conditions under which the system is evaluated should be at least as
rigorous as the conditions specified in the corresponding EPA test procedure.
For example, in the case of tank or pipeline tightness testing, the test should
include a temperature difference between the delivered product and that
already present in the tank or pipeline.
4. The evaluation results must contain the same information as the EPA
standard results sheet and should be reported according to the same general
format.
5. The evaluation must include physical testing of a full-sized version of the leak
detection system, and a full disclosure must be made of the experimental
conditions under which the evaluation was performed, and the conditions
under which its use is recommended. An evaluation based solely on theory
or calculation is not sufficient.
VI
-------
ABSTRACT
This report presents a standard test procedure for evaluating the performance of leak
detection systems for use in the pipelines associated with underground storage tanks.
The test procedure is designed to evaluate these systems against the performance
standards in EPA's underground storage tank regulations (40 CFR Part 280, Subpart D),
which cover an hourly test, a monthly monitoring test, and a line tightness test. The test
procedure can be used to evaluate any type of system that is attached to the pipeline
and monitors or measures either flow rate or changes in pressure or product volume.
This procedure can be used to evaluate a leak detection system that can relate the
measured output quantity to leak rate (in terms of gallons per hour) and systems that
use an automatic preset threshold switch. The test procedure can evaluate systems
used to test pressurized pipelines or suction pipelines that are pressurized for the test.
The test procedure offers five options for collecting the data required to calculate
performance. The results of the evaluation are reported in a standard format on forms
provided in the appendices of the report.
This report was submitted in fulfillment of Contract No. 68-03-3409 by Vista Research,
Inc., under the sponsorship of the U.S. Environmental Protection Agency. This report
covers a period from March 1989 to March 1990, and work was completed as of July
1990.
VII
-------
TABLE OF CONTENTS
Disclaimer i
Foreword ii
Preface iii
Abstract vii
List of Figures xi
List of Tables xii
Acknowledgements xiii
Section 1: I ntroduction 1
1.1 Types of Systems Covered by this Protocol 2
1.1.1 Summary of the EPA Regulation for Pressurized Pipelines 3
1.1.2 Interpretation of the Regulation 3
1.2 Objective 4
1.3 For Whom Was This Report Prepared? 5
1.4 Safety 6
1.5 Getting Started 6
1.6 Units 8
1.7 Report Organization 9
1.8 Notification of Protocol Changes 11
Section 2: Performance 13
2.1 Definition of a Leak 13
2.2 Definition of Performance 14
Section 3: General Features of the Evaluation Protocol 21
3.1 Pipeline Configuration 21
3.2 Summary of Options for Estimating Performance with this Protocol 23
3.2.1 Generating the Noise Histogram 24
3.2.2 Generating the Signal-plus-noise Histogram 26
3.2.3 Generating Histograms with Leak Detection Systems that Use a
Multiple-test Strategy 28
3.3 Conducting the Evaluation 29
3.4 Accuracy of the Evaluation 32
3.5 Other Acceptable Evaluation Protocols 34
Section 4: Equipment Needed for Generating Evaluation Conditions 35
4.1 Line Pressure 35
4.1.1 Equipment and Instrumentation for Generating Line Pressure 36
4.1.2 Measurement of Line Pressure 36
VIII
-------
4.2 Leak Rate 36
4.2.1 Equipment and Instrumentation for Generating Leaks 38
4.2.2 Measurement of Leak Rate 39
4.2.3 Relationship Between the Signal and the Noise 40
4.3 Pipeline Compressibility Characteristics 42
4.3.1 Equipment and Instrumentation for Modifying Pipeline
Compressibility 43
4.3.2 Measurement of Pipeline Compressibility 44
4.4 Product Temperature 47
4.4.1 Equipment and Instrumentation for Generating Product
Temperature 49
4.4.2 Measurement of Product and Ground Temperatures 52
4.5 Trapped Vapor 53
4.5.1 Equipment and Instrumentation for Generating Trapped Vapor 53
4.5.2 Measurement of Trapped Vapor 55
Section 5: Selection of Evaluation Conditions 57
5.1 Temperature Conditions in the Pipeline 57
5.2 Induced Leak Rates 62
5.2.1 Known Test Conditions 63
5.2.2 Procedures for Blind Testing 65
5.2.2.1 Procedure 1 66
5.2.2.2 Procedure 2 67
Section 6: Evaluation Procedure for Systems that Report a Flow Rate 70
6.1 Performance Characteristics of the Instrumentation 70
6.2 Development of the Noise and the Signal-Plus-Noise Data 71
6.3 Evaluation Procedure 72
6.3.1 Option 1 - Collect Data at a Special Pipeline Test Facility 72
6.3.2 Option 2 - Collect Data at One or More Instrumented Operational
LIST Facilities 77
6.3.3 Option 3 - Collect Data over a 6- to 12-month Period at 5 or More
Operational LIST Facilities 77
6.3.4 Option 4 - Collect Data over a 6- to 12-month Period at 10 or More
Operational LIST Facilities 82
6.3.5 Option 5 - Develop the Noise and Signal-plus-noise Data from an
Experimentally Validated Computer Simulation 84
6.4 Calculation of PD and PFA 87
Section 7: Evaluation Procedure for Systems that Use a Preset Threshold 91
7.1 Performance Characteristics of the Instrumentation 91
7.2 Development of the Noise and the Signal-Plus-Noise Data 92
IX
-------
7.3 Evaluation Procedure 93
7.4 Calculation of PD and PFA 94
Section 8: Detection Tests with Trapped Vapor in the Pipeline 96
Section 9: Reporting of Results 97
Section 10: Technical Basis for Values Used in the Protocol 99
10.1 Range of Temperature Conditions 99
10.2 Number of Tests 99
10.3 Range of the Bulk Modulus 102
10.4 Vapor Pockets 102
References 103
Appendix A: Form to Present a Description of the Pipeline Leak Detection System
Evaluated According to the EPA Test Procedure 104
Appendix B: Form to Present the Results of the Performance Evaluation Conducted
According to the EPA Test Procedure 120
Appendix C: Protocol Notification Form 153
Appendix D: Random Selection of Leak Rates 155
Appendix E: Statistics 157
-------
LIST OF FIGURES
Figure 1.1 Summary of the Report Organization 10
Figure 2.1 Histogram (a) and frequency distribution (b) of the noise compiled from
25 leak detection tests on nonleaking pipelines for a volumetric leak
detection system 17
Figure 2.2 Cumulative frequency distribution of the noise derived from the
frequency distribution in Figure 2.1 18
Figure 2.3 Cumulative frequency distribution of the signal-plus-noise generated for
a leak rate (i.e., signal) of -0.10 gal/h using the cumulative frequency
distribution of the noise shown in Figure 2.2 18
Figure 2.4 Statistical model for calculating the PD and PFA of a pipeline leak
detection system 19
Figure 3.1 Pressure-volume relationship for a 2-in.-diameter, 200-ft steel pipeline
with and without vapor trapped in the pipeline system 26
Figure 4.1 Schematic diagram of an apparatus to generate small and large leaks in
the pipeline 39
Figure 4.2 Mechanical device to modify the compressibility characteristics of the
pipeline system 43
Figure 4.3 Pressure-volume relationship for a 2-in.-diameter, 165-ft pipeline (a)
without and (b) with a mechanical line leak detector 45
Figure 4.4 Pressure-volume relationship for a 2-in.-diameter, 200-ft steel pipeline
when the compressibility device is attached to the line and when it is
not 46
Figure 4.5 Product temperature changes predicted for different dispensing
operations using a heat transfer model: (a) temperature of the backfill
and soil is constant, (b) temperature of the backfill and soil that is
produced by circulating product through the pipeline for 16 h at a
temperature that was initially constant and 9°F higher than the backfill
and soil, (c) time history of the product temperature changes in the
pipeline for the initial ground conditions shown in (a) and (b) 50
Figure 4.6 Geometry of the temperature measurements to be made in the backfill
and soil surrounding an underground pipeline 51
Figure 4.7 Mechanical device for trapping vapor in a pipeline system 54
Figure 5.1 Model predictions of the temperature changes that occur in a pipeline
(a) after a 1-h circulation period and (b) after a 5-min circulation period
for an initial temperature difference between the product circulated
through the pipeline and the backfill and soil of 4.5°F 59
XI
-------
LIST OF TABLES
Table 1.1 Equivalent Leak Rates 4
Table 2.1 Estimate of the Average Monthly Loss of Product from an Undetected
Leak in a Pipeline 14
Table 4.1 Recommendations for Measuring Leak Rate 40
Table 4.2 Volume of Trapped Vapor in a Tube 1.5 in. in Diameter and 3.5 in. in
Length as a Function of Pipeline Pressure 55
Table 5.1 Number of Tests Required for Each Range of Temperature Conditions ...57
Table 5.2 Recommended Procedure for Generating a Temperature Condition at
an Instrumented Test Facility 61
Table 5.3 Example of Test Conditions When More Than One Test Can Be Done
for a Temperature Condition 64
Table 5.4 Recommended Leak Rates for Procedure 2 67
Table 5.5 Illustration of a Possible Test Matrix for Evaluation of a Leak Detection
System at 0.1 gal/h 67
Table 6.1 Values of the Cumulative Frequency Distribution of the Noise Shown in
Figure 2.2 88
Table 6.2 Values of the Cumulative Frequency Distribution of the Signal-plus-
noise Shown in Figure 2.3 Generated for Leak Rate (i.e., Signal) of 0.10
gal/h 89
Table 7.1 Values of the Cumulative Frequency Distribution of the Noise Shown in
Figure 2.2 93
Table 7.2 Values of the Cumulative Frequency Distribution of the Signal-plus-
noise Shown in Figure 2.3 Generated for Leak Rate (i.e. Signal) of 0.10
gal/h 94
Table 10.1 Experimental Uncertainty on the Standard Deviation of the Noise and
Signal-plus-noise Histograms, the Smallest Leak Rates that Can Be
Detected with a PD of 0.95 and a PFA of 0.05, and the PD and PFA
Characterized by the 95% Confidence Intervals on the Standard
Deviation of Detection of a Leak Rate of 0.10 gal/h 100
Table 10.2 Experimental Uncertainty on the Standard Deviation of the Noise and
Signal-plus-noise Histograms, the Smallest Leak Rates that Can Be
Detected with a PD of 0.95 and a PFA of 0.05, and the PD and PFA
Characterized by the 95% Confidence Intervals on the Standard
Deviation of Detection of a Leak Rate of 0.20 gal/h 100
Table 10.3 Experimental Uncertainty on the Standard Deviation of the Noise and
Signal-plus-noise Histograms, the Smallest Leak Rates that Can Be
Detected with a PD of 0.95 and a PFA of 0.05, and the PD and PFA
Characterized by the 95% Confidence Intervals on the Standard
Deviation of Detection of a Leak Rate of 3.0 gal/h 101
XII
-------
ACKNOWLEDGMENTS
This protocol for evaluating underground storage tank pipeline leak detection systems
was prepared by Joseph W. Maresca, Jr., Robert M. Smedfjeld, Richard F. Wise, and
James W. Starr for the U.S. Environmental Protection Agency's (EPA 's) Risk Reduction
Engineering Laboratory (RREL) on Work Assignment 18 of EPA Contract No. 68-03-
3409. Anthony N. Tafuri was the Technical Program Monitor on the Work Assignment
for EPA/RREL. Technical assistance and review were provided by Thomas Young and
David O 'Brien of the EPA's Office of Underground Storage Tanks (OUST). Over 50
copies of the first draft of this protocol were distributed for external technical review to
petroleum industry trade associations, manufacturers of pipeline leak detection systems,
regulatory agencies, and owners/operators of underground storage tank systems. Many
of these organizations specifically requested to participate in the review and generously
offered their comments and suggestions. A second draft of this protocol underwent
technical review by members of the manufacturing, user, and regulating communities
attending an EPA-sponsored workshop held in Kansas City, Missouri, in March 1990.
This document was edited by Monique Seibel, who also prepared the technical
illustrations. Pamela Webster prepared the document for publication.
XIII
-------
SECTION 1
INTRODUCTION
A protocol has been developed that can be used to evaluate the performance of leak
detection systems or methods used to test the integrity of pipelines associated with
underground storage tanks (USTs). The protocol applies to leak detection systems or
methods that are physically attached to the pipeline and can relate the measured output
quantity to a leak rate associated with the loss of product through a hole in a pipeline
under pressure. The system does not, however, have to be one that reports a quantified
leak rate. For example, systems that use an automatic preset threshold switch can also
be evaluated with this protocol. The performance results are reported in terms of leak
rate (in gallons per hour), probability of detection (PD), and probability of false alarm
(PFA). The protocol specifically addresses the performance of these leak detection
systems for the leak rates, PD, and PFA specified in the technical standards prescribed in
the United States Environmental Protection Agency (EPA) UST regulations (40 CFR
Part 280 Subpart D) [1]. The protocol specifically covers all of the internal EPA release
detection options for piping, but does not cover the external leak detection options (those
for vapor and groundwater monitoring). A separate protocol has been developed for
these external systems [2,3]. Common types of leak detection systems that can be
evaluated with this protocol include systems that measure pressure, volume, or flow-rate
changes in the pipeline. This protocol addresses both pressurized and suction piping
systems and assumes that if release detection is required for a suction system, the line
will be pressurized for the test.
The protocol is flexible enough to permit a wide range of approaches to collecting the
test data necessary to perform the evaluation and yet is specific enough for the results of
each approach to be repeatable. The data needed to perform the evaluation can be
collected either at a special test facility or at one or more operational UST facilities, such
as retail stations or industrial storage sites. The same protocol can be used for an
hourly test, a monthly monitoring test, and a line tightness test.
Because pressurized pipelines present the potential for a large release of product if a
leak occurs, the EPA regulation requires stringent and frequent testing. Methods of
release detection for pressurized UST pipelines must handle two different but equally
important leak scenarios. In the first scenario, a large release occurs over a short time.
The submersible pump that brings product through the pipeline system can pressurize
the line for product to be dispensed even though there may be a large hole or fissure in
the line. When the line is under pressure, much product can be lost in a short time. In
the second scenario, small amounts of product are released over a long period of time; if
the leak continues undetected, the net loss of product can be as great as in the first
scenario. The EPA regulations for pressurized pipelines require that both leak detection
scenarios be addressed. In some instances, the same leak detection system can be
-------
used to address each scenario; however, the test procedure, the analysis, and the
criterion used to detect a leak may differ. The first scenario requires a test that can be
conducted quickly and frequently and that can be used to detect the presence of large
leaks having the potential to cause serious environmental damage over a period of tens
of minutes to several hours. The second scenario requires a periodic precision test* that
can be used to detect the presence of very small leaks having the potential to cause
serious environmental damage over a period of a month to a year. The protocol
described in this report can be used to evaluate the performance of systems designed to
handle each scenario.
The EPA regulation states that "suction piping appears to be intrinsically much safer
(than pressurized piping) because product is transferred at less than atmospheric
pressure by a pump near the dispenser drawing product from the tank by suction, and
failures will result in air or groundwater flowing into the pipe rather than product being
released during operation" [1]. As a consequence, the release detection requirements
for suction piping presented in the regulation are significantly less stringent than those
for pressurized piping. Suction piping is exempt from release detection requirements if
the "suction piping meets six design and operating standards concerning pressure,
slope, run of the piping system, and use of properly located check valves" [1]. If these
six standards are not met, the suction piping system must be tested with one of the
monthly monitoring options or must be tested once every three years with a line
tightness test. One method of testing a suction piping system is to isolate the line from
the tank, pressurize it, and use one of the systems designed for pressurized lines.
It is important to note that in this protocol performance estimates are made in such a
way that they can be compared to the technical standards prescribed in the EPA
regulation. It should be assumed that the manufacturer will use the best equipment and
the best operators (if operators are required) available at the time of the evaluation. The
evaluation is not designed to determine the functionality of the system (i.e., whether it
operates as intended), nor is it meant to assess either the operational aspects of the
system (e.g., the adequacy of the maintenance and calibration procedures) or the
robustness of the system.
1.1 TYPES OF SYSTEMS COVERED BY THIS PROTOCOL
Leak detection systems for both pressurized and suction piping can be evaluated
with this protocol. The release detection requirements for this piping are described in
Sections 280.40, 280.41 (b), 280.43(h), and 280.44(a)-(c) of the EPA underground
storage tank regulation [1]. The protocol does not specifically include a methodology
A precision test, as used in this protocol, refers to any system that can detect a leak of 0.2 gal/h
or better (required for monthly monitoring tests) or a leak of 0.1 gal/h or better (required for line
tightness tests).
-------
for evaluating vapor and groundwater monitoring described in Sections 280.43(e)
and (f); as indicated above, separate protocols have been developed for evaluating
these types of systems [2,3].
1.1.1 Summary of the EPA Regulation for Pressurized Pipelines
The EPA regulation requires two types of leak detection tests for underground
pressurized piping containing petroleum fuels. First, as stated in Sections
280.4l(b)(1)(i) and 280.44(a), underground piping must be equipped with an
automatic line leak detector that will alert the operator to the presence of a leak
by restricting or shutting off the flow of the regulated substance through the
piping or by triggering an auditory or visual alarm. The automatic line leak
detector must be capable of detecting leaks of 3 gal/h defined at a line pressure
of 10 psi within an hour of the occurrence of a leak with a PD of 95% (0.95) and a
PFA of 5% (0.05). The test is designed to detect the presence of very large leaks
that may occur between regularly scheduled checks with the more accurate
monthly monitoring tests or annual line tightness tests.
Second, the regulation also requires either an annual line tightness test or one of
four monthly monitoring tests. The annual line tightness test must be capable of
detecting a leak as small as 0.1 gal/h (defined at a pressure which is 150% of the
operating pressure of the line) with a PD of 95% and a PFA of 5%. One of the
monthly methods allowed is a line test that can detect leaks as small as 0.2 gal/h
(defined at the operating pressure of the line) with a PD of 95% and a PFA of 5%.
This option, which is allowed by Section 280.44(c) and described under Other
Methods that Meet a Performance Standard in Section 280.43(h) of the
regulation, requires that the performance of the method be quantified. This
quantitative option covers the use of any type of pipeline leak detection system
(line pressure monitor, automatic shutdown line leak detector, etc.) that conducts
a precision test on the pipeline system and that can satisfy the performance
requirements. The monthly monitoring requirement may also be met by one of
three other methods of leak detection: vapor monitoring, groundwater
monitoring, or interstitial monitoring. The regulation lists specific requirements
that each of these three methods must meet. These requirements are designed
to assess whether the method is applicable to the local backfill, groundwater, and
soil conditions. In general, an engineering evaluation of the site is required
whenever a method of leak detection that is external to the tank is used.
1.1.2 Interpretation of the Regulation
The standard for automatic line leak detectors (Section 280.44(a)) requires that a
leak of 3 gal/h or larger (defined at 10 psi with a PD of 95% and a PFA of 5%)
must be detected within one hour of its occurrence. This suggests that a test of
-------
the line must be conducted once per hour or that the leak detector must be able
to sense a leak of this magnitude within one hour of its occurrence.
The automatic line leak detection section of the regulation (Section 280.44(a))
was intended to allow the use of mechanical line leak detectors [4]. Thus, the
performance specification in the regulation is identical to the performance claim
made by the manufacturers of this type of system. However, this regulatory
standard does not preclude the use of other types of automatic systems as long
as they can conduct at least one test per hour and detect a release of 3 gal/h
(defined at 10 psi with a PD of 95% and a PFA of 5%); for example, a line pressure
monitoring system that has the required performance can also be used.
The regulation also allows automatic line leak detectors to be used for precision
testing, provided that the detection systems' performance meets either the
monthly monitoring test requirements in Sections 280.44(c) and 280.43(h) - (i) or
the annual precision test requirements in Section 280.44(b).
The regulation specifies the minimum leak that a system must be able to detect
at specific pressures. Since leak rate varies as a function of pressure, the leak
detection test can be conducted at different pressures provided that the
determinable leak rate at the specified test pressure is equivalent to or more
stringent than the one mandated in the regulation. Examples of equivalent leak
rates are given in Table 1.1. (They were calculated from Eq. (4.1), described in
Section 4.2.)
Table 1.1. Equivalent Leak Rates
Leak
Rate
(gal/h)
3
0.1
0.2
Test
Pressure
(psi)
10
45
30
Equivalent
Leak Rate*
(gal/h)
4.25
0.07
0.16
Equivalent Test
Pressure
(psi)
20
20
20
* Based on a theoretical calculation which assumes that turbulent flow occurs through
a sharp-edged orifice
1.2 OBJECTIVE
The objective of this protocol is to provide a standard procedure for evaluating the
performance of leak detectors that monitor or test the piping associated with
underground storage tanks. The type of detector addressed by this protocol is
located on a single pipeline connecting the tank with the dispenser. Both
pressurized- and suction-piping leak detection systems are included; however,
suction pipelines must be pressurized for a test. The protocol can be used to
evaluate any leak detection system that can relate the measured output quantity to
-------
leak rate (in terms of gallons per hour); systems that use an automatic preset
threshold switch can also be evaluated with this protocol. Interstitial leak detection
systems can be evaluated with a variation of this protocol, but it should be noted that
the protocol is not specifically designed for these systems.
This protocol can be used to evaluate two types of pipeline leak detectors: (1) those
that perform hourly tests of the line and that claim to detect leak rates of 3 gal/h
defined at 10 psi with a PD of 0.95 and a PFA of 0.05, and (2) those that perform
either a monthly monitoring test with a claimed performance of 0.2 gal/h or a line
tightness test (annually for pressurized piping or every 3 years for suction piping)
with a claimed performance of 0.1 gal/h with a PD of 0.95 and a PFA of 0.05. All
pipeline leak detection systems will be evaluated for accuracy and reliability for a
specified pipeline configuration, under a wide range of ambient test conditions
(primarily product temperature), and, at a minimum, at the leak rate specified in the
EPA regulation. The probability of false alarm will be estimated at the threshold used
by the manufacturer, and the probability of detection will be estimated at the leak rate
specified in the EPA regulation.
With one slight difference, the same procedure will be used to evaluate the
performance of the monthly monitoring test, the annual line tightness test, and the
hourly test. For the monthly monitoring test, the probability of detection will be
estimated at a leak rate of approximately 0.2 gal/h, while for the line tightness test
the probability of detection will be estimated at a leak rate of approximately 0.1 gal/h;
a 3-gal/h leak will be used in the hourly test. The evaluation procedure requires that
the performance characteristics of the instrumentation be estimated and that the
performance in terms of leak rate, PD, and PFA be determined for the specified
pipeline configuration and a wide range of product temperature conditions. Any
automatic line leak detector that can address the 3-gal/h standard will be evaluated
under the same range of environmental and pipeline-configuration conditions as the
systems that conduct monthly monitoring and line tightness tests. The protocol
requires that the operator or system controller calculate and report both the PFA at
the manufacturer's threshold and the PD for the appropriate leak rate specified in the
EPA regulation. If it has sufficient performance, an automatic line leak detector used
to satisfy the hourly test can also be used to satisfy the monthly monitoring test or
the annual line tightness test.
1.3 FOR WHOM WAS THIS REPORT PREPARED?
This report is intended for any person, group, or organization that wants to evaluate
a pipeline leak detection system designed to meet one or more aspects of the EPA
regulation, and that may in addition want to report the results of such an evaluation.
Two groups that will find the report useful are manufacturers of pipeline leak
detection systems and third-party evaluators of such systems. Although not
-------
specifically intended for regulators or owners and operators of underground storage
tank systems, it may nonetheless provide these groups with useful information
regarding the requirements for evaluation.
1.4 SAFETY
This protocol does not address the safety considerations involved in evaluating leak
detection systems for pipelines containing petroleum products. It is, however,
imperative that the leak detection system and the evaluation equipment and facilities
be safe and be used safely. Whether the leak detection system is to be evaluated at
one or more operational LIST facilities or at a special test facility, the organization
supplying the leak detection system should provide a standard safety procedure for
operating the system and should explain this procedure to the organization doing the
evaluation. Similarly, the organization doing the evaluation should provide a
standard safety procedure for the use and handling of the evaluation equipment, the
pipeline and storage tank facilities, and the product in the pipeline and tank system
and should explain this procedure and how to use safety equipment such as fire
extinguishers to the organization whose detection system is being evaluated. This
should be done before any testing begins. All local, state, and federal health, safety,
and fire codes and regulations should be adhered to; these codes and regulations
take precedence if there is any conflict between them and the instructions in this
document.
1.5 GETTING STARTED
One should read this document in its entirety before attempting to evaluate a pipeline
leak detection system. Having done this, one should determine how the evaluation
will be conducted and prepare a detailed operational procedure. This is particularly
important because this protocol could have been prepared as six separate
documents to evaluate the six different types of pipeline leak detection systems
covered by this protocol. The particulars of the evaluation procedure depend on
which performance standard the system will be evaluated against (i.e., hourly test at
3 gal/h, monthly monitoring test at 0.2 gal/h, or line tightness test at 0.1 gal/h) and
whether the leak detection system measures the flow rate and uses it to determine
whether the pipeline is leaking, or uses an automatic preset threshold switch and
does not directly measure and report flow rate.
There are a number of important choices that the evaluator must make to conduct
the evaluation. There are five options for collecting data: (1) at a special
instrumented test facility, (2) at one or more instrumented operational LIST facilities,
(3) at five noninstrumented operational LIST facilities where pipeline integrity has
been verified, (4) at ten or more noninstrumented operational LIST facilities where
the status of the pipeline is unknown, or (5) by means of an experimentally validated
-------
computer simulation. Of these five, the first four are the most common. The option
selected depends on the time and facilities available for the evaluation. The protocol
requires that the data be collected on one or more pipeline systems which satisfy a
specific set of minimum characteristics established by this protocol, over a very wide
range of product temperature conditions representative of those found throughout
the United States in all four climatic seasons, and for one or more leak rates that are
defined by the EPA regulations; the protocol also provides a means to verify that all
of these conditions are satisfied.
Another choice the evaluator must make is whether the test crew and/or the
organization supplying the system will have full knowledge of the test conditions
beforehand or whether they will be placed in a blind testing situation. In either case,
a test matrix of temperature and leak conditions must be defined and data must be
collected according to this matrix. The protocol provides a way to develop a test
matrix for each type of condition. The protocol is designed to minimize any
advantages that the test crew might have because of its familiarity with the tests
conditions. Thus, the performance estimates should be identical regardless of
whether the test conditions were known a priori. Two blind testing techniques are
provided that can be most easily implemented at an instrumented test facility; blind
testing, it should be noted, takes more time and effort to complete.
Before the evaluation is begun, the vendor must describe the important features of
the leak detection system to be evaluated; for this purpose summary sheets are
included in Appendix B. Once the system has been defined, the data needed to
perform the evaluation can be collected. Three types of measurements must be
made. First, the performance characteristics of each instrument that is part of the
system must be determined. (This means, for example, the resolution, precision,
accuracy, and dynamic range of instruments such as pressure sensors and
temperature sensors.) This ensures that the instruments are functioning properly.
Second, the data with which to make an estimate of performance in terms of leak
rate, probability of detection, and probability of false alarm must be collected. This is
the heart of the evaluation, and much of this report focuses on how to collect and
analyze these data. This protocol requires that a minimum of 25 leak detection tests
on a nonleaking line be conducted over a wide range of pipeline temperature
conditions. Justification for requiring 25 tests is presented in Section 10 of this
report. Additional tests during which a leak is generated in the pipeline system are
also necessary. The protocol is designed to use the leak rate specified in the
appropriate EPA regulatory standard. Third, the sensitivity of the leak detection
system to the presence of small quantities of vapor trapped in the pipeline system
must be determined. Only a few tests are required to assess this sensitivity,
because a simple field measurement technique is provided that can be used prior to
testing to determine whether or not a pipeline contains any trapped vapor. Once
-------
these data have been collected, the analysis and reporting procedures are relatively
straightforward. The results of the evaluation are to be reported on the form
provided in Appendix A. Seven attachments to the evaluation form are provided for
describing the system that was evaluated; these can be found in Appendix B.
The protocol specifies certain equipment, apparatuses, and measurement systems
to be used in the evaluation. None of these are particularly complex or
sophisticated, and a description of each is provided. The protocol allows for the use
of other equipment not specified by this protocol provided it has the same
functionality and performance as the equipment described.
Only a limited knowledge of mathematics is required to implement this standard test
procedure. All of the mathematics can be performed with a calculator or one of the
many spreadsheets available on personal computers. This protocol requires that the
evaluator be able to:
• sort data from the smallest value to the largest value
• calculate the mean and standard deviation
• fit a regression line to a set of data
• use a random number generator or draw random numbers from a container
• plot and read an x-y graph or be able to linearly interpolate between numbers
in a table
The formula for calculating the mean and standard deviation and for calculating the
regression line to a set of data is summarized in Appendix E.
1.6 UNITS
In this report, the most common quantities are length, volume, time, flow rate,
temperature, and pressure. In accordance with the common practice of the leak
detection industry , these quantities are presented in English units, with the
exception of small volumes measured in a graduated cylinder, in which case the
metric units are used and the English units are presented in parentheses. Length is
measured in inches (in.) and feet (ft). Large volumes are measured in gallons (gal);
small volumes, which are the exception, are measured in milliliters (ml). Time is
measured in units of seconds (s), minutes (min), and hours (h). All flow rate
measurements made in this report are calculated from measurements of volume and
time; flow rate quantities are presented in gallons per hour (gal/h), although the
measurements necessary to calculate flow rate will generally be made in units of
volume (ml or gal) and units of time (s, min, h) and must be converted. Pressure is
measured in units of pounds per square inch (psi). Finally, temperature quantities
8
-------
are measured in degrees Fahrenheit (°F), although some temperature measurement
systems used in the leak detection industry employ degrees Centigrade (°C).
1.7 REPORT ORGANIZATION
This report is organized in such a way that it facilitates the evaluation of many
different types of leak detection systems against different performance standards
and allows the evaluator great flexibility in the approach used for generating the data
required to estimate the performance of the system. The report organization is
summarized in Figure 1.1. The reason for organizing the report in this way is to
make it easier for the evaluator to identify the steps for completing an evaluation
(which are presented in Sections 6 and 7) without being encumbered by too much
detail. Relevant details are provided in other sections.
Section 1 introduces the protocol for evaluating pipeline leak detection systems.
Section 2 describes the standard procedure for evaluating the performance of any
leak detection system in terms of leak rate, probability of detection, and probability of
false alarm. As part of the evaluation two histograms are developed: one of the
noise that is present during tests on a nonleaking pipeline and the other of the signal-
plus-noise during tests on a leaking pipeline. The EPA regulation specifies that
certain leak detection systems must be able to detect certain flow rates defined at
prescribed line pressures. The flow rate of the leak generated for the signal-plus-
noise histogram will therefore be appropriate for the type of system being evaluated
(0.1 gal/h for line tightness testing systems, 0.2 gal/h for monthly monitoring
systems, and 3.0 gal/h for hourly testing systems) and will be referred to in this report
as the EPA-specified leak rate.
Section 3 gives a brief overview of the evaluation procedure that is used to derive the
performance estimate. The accuracy of the evaluation procedure and how to assure
the integrity of the evaluation are discussed in Section 3.4; the use of other methods
of evaluation is discussed in Section 3.5.
-------
—J
^ Wifsste^
i « s F-AK out- CT ION s fvrtvs
COVERED aYTH,3 PROTOCOL
» OBJECTIVES OF Tut PROTOCOL
I « PS«.V.B*-< ti>Wr
-------
The equipment needed to conduct the evaluation is described in Section 4, including
the sensor system and the requirements for temperature and pressure sensors.
Section 4 includes a general description of the apparatus required to induce and
measure a leak in the pipeline and the various devices needed to characterize the
temperature condition of the product in the pipeline, generate a known volume of
trapped vapor in the line, and adjust the compressibility of the pipeline system.
Section 4 also describes the procedures for making measurements with this
equipment. All of the equipment can be assembled with simple mechanical parts.
All the equipment can be mounted at existing inlets or outlets so that no new
openings in the pipeline are necessary.
Section 5 describes two approaches to selecting and defining the temperature and
leak conditions required to conduct the evaluation. In the first approach, the leak
rate, temperature condition, vapor pocket, and compressibility characteristics of the
pipeline are known by the testing crew before each leak detection test. In the
second approach, the test conditions are not known until all the tests have been
completed. Both approaches are equally acceptable and will result in identical
performance estimates.
Section 6 describes the evaluation procedure for systems that report a flow rate, and
Section 7 describes the procedure for systems that use a preset threshold. There
are five options for collecting the noise and signal-plus-noise data that are required
for the performance calculations. In Sections 6 and 7, a separate procedure is
provided for each of these five options. Sample calculations on how to estimate the
probability of detection and probability of false alarm are also included.
Section 8 describes how to determine the sensitivity of a pipeline leak detection
system to vapor that may be trapped in the line.
Section 9 describes the minimum information required to describe the leak detection
system and how to tabulate and report the results of the evaluation.
Section 10 presents the technical basis for the selection of the test conditions.
1.8 NOTIFICATION OF PROTOCOL CHANGES
A draft of this protocol was reviewed by regulators, manufacturers of pipeline leak
detectors, providers of pipeline leak detection services, evaluators of leak detection
equipment, scientists and consulting engineers, and owners/operators of
underground storage tank systems. While the approach used in this protocol has
been used to evaluate the performance of underground storage tanks, it has not
been widely used for pipelines. Since clarification or modification of the procedures
in this protocol may be required once the protocol is implemented by the industry, the
11
-------
EPA requests that any user of the protocol fill out the notification form in Appendix C
and mail it to the EPA at the following address:
Office of Underground Storage Tanks
U.S. Environmental Protection Agency
Attention: Pipeline Evaluation Test Procedure
401 M Street, S. W.
Mail Stop OS-410
Washington, D. C. 20460
This will place users on a mailing list so that they can be notified of any changes to
the protocol. Comments or suggestions on how to improve the protocol are also
welcomed and should be addressed to the same agency.
12
-------
SECTION 2
PERFORMANCE
To understand how the evaluation is conducted, it is necessary to know the definition of
a leak, the definition of performance in terms of probability of detection and probability of
false alarm, and how performance is estimated. It should be noted that pipeline
configuration and ambient conditions can influence the evaluation.
2.1 DEFINITION OF A LEAK
The flow rate produced by a leak in the pipeline will change with line pressure,
increasing when pressure is high and decreasing when pressure is low. The total
volume of product that can be lost from a leak in a pipeline is the sum of (1) the
volume of fluid lost when product is being dispensed and (2) the volume of fluid lost
when product is not being dispensed. The total volume of product lost during
dispensing is estimated by multiplying the leak rate (defined at the operating
pressure of the line) by the duration of the dispensing. Even small holes may result
in a release of product at a rate of several gallons per hour. The volume of product
lost in the intervals between dispensing is more difficult to estimate accurately.
Unless the hole in the line is excessively large, the total volume that is typically
released from a leaking pipeline when no dispensing is occurring ranges from 0.03 to
0.06 gal. Product is released between dispensing periods because the pipeline
system is elastic, and, under pressure, it expands. At the operating pressures
typically found at retail stations, the pipeline system expands 0.03 to 0.06 gal. As the
pressure decreases, product is released through the hole at a decreasing rate. Once
the pressure reaches zero, no further product is lost. If the hole is very small, the
leak may stop before the pressure reaches zero; if the hole is very large, the entire
contents of the line may be released.
The values in Table 2.1 illustrate the average monthly release of product resulting
from a missed detection, given that product was dispensed at a rate of 5 gal/min to a
known number of cars each requiring 10 gal of fuel. The loss of product was
calculated on the basis of leaks of 0.1, 0.2, and 10 gal/h, and the averaged missed
detections were assumed to be leaks with flow rates that were 50% of these values.
It was further assumed that the product was dispensed at 30 psi, that the average
volume of product lost in the intervals between dispensing was 0.0264 gal, and that
the time between dispensing was long enough for this volume of product to be
released. When the leak is small, the quantity of the product released during
dispensing is also small relative to the quantity of product released during the
intervals when product is not being dispensed. The reverse is true if the leak is
large. An average hourly leak rate of 0.1 gal is equivalent to a release of 72 gal per
month.
13
-------
Table 2.1. Estimate of the Average Monthly Loss of Product from an Undetected Leak in a
Pipeline
Leak
Rate
(gal/h)
0.1
0.1
0.2
0.2
10.0
10.0
Average
Missed
Detection
(gal/h)
0.05
0.05
0.1
0.1
5.0
5.0
Monthly
Throughput
(gal)
16,000
50,000
16,000
50,000
16,000
50,000
Number
of Cars
1,650
5,000
1,650
5,000
1,650
5,000
Product
Lost
While
Dispensing
(gal)
3
8
5
16
267
833
Product
Lost
While Not
Dispensing
(gal)
42
132
42
132
42
132
Total Monthly
Liability
(gal)
45
140
47
148
309
965
(gal/h)
0.06
0.19
0.07
0.21
0.43
1.34
2.2
DEFINITION OF PERFORMANCE
A complete specification of system performance requires a description of the
probability of false alarm (PFA) and the probability of detection (PD) at a defined leak
rate, LR, and an estimate of the uncertainty of the PD and PFA- These estimates
should be made over the range of conditions under which the system will actually be
used. They can be made from a performance model based on the histograms of the
noise and the signal-plus-noise. The actual calculations will be made with another
representation of the histogram called the cumulative frequency distribution.
The probability of detection is defined as the number of leaking pipelines that a
system would detect if all the pipelines tested were leaking. The probability of
detection is expressed as a decimal fraction or a percentage. Thus, a probability of
detection of 95%, which may also be written as 0.95, would suggest that the system
will correctly declare leaks in 95% of the leaking pipelines tested. Missed
detections occur if the system fails to declare a leak when one is present; this
occurs most frequently when the leak is small compared to the background noise
(i.e., the pressure fluctuations that occur in nonleaking pipeline systems, due, for
example, to thermal expansion and contraction of the product). The probability of a
missed detection (i.e., a false negative) is directly related to the probability of
detection. If the probability of detection is 95%, then the probability of missed
detection is 5%; if the probability of detection is 99.9%, then the probability of missed
detection is 0.1%. The probabilities of detection and/or missed detection are
estimated from the cumulative frequency distribution of the signal-plus-noise.
The probability of false alarm (i.e., a false positive) is defined as the number of
tight (nonleaking) pipelines that a system would declare leaking if all the pipelines
14
-------
tested were tight. Thus, a probability of false alarm of 5% would suggest that the
system will incorrectly declare leaks in 5% of the nonleaking pipelines tested. The
probability of false alarm can be estimated from the cumulative frequency distribution
of the noise once a threshold has been selected.
Detection of leaks in pipeline systems is an example of the classical statistical
problem of finding a signal in the presence of noise. In this case, the signal is the
flow rate of the liquid through a hole in the pipeline, defined at a constant pressure.
Note that the primary measurement of the leak detection system may be pressure, or
it may be volume, or it may be something entirely different; but it is the leak rate that
is the quantity of interest, i.e., the signal. The term noise refers to the amount of
fluctuation that occurs in the absence of the signal. Thus, in order to assess the
performance of a leak detection system, one must know the fluctuation level of the
measured quantity both with and without the presence of the signal. Noise
represents effects that would be misinterpreted (by the particular leak detection
system) as a leak when no leak was present or that would mask an existing leak.
These are effects that have characteristics similar to those of a leak. Note that the
noise is a system-specific quantity. If the leak detection system attempts to detect
the presence of the signal (leak) by measuring the pressure drop associated with
flow out of the line, then any physical mechanism that produces pressure changes in
a nonleaking line and that looks like the pressure changes produced by a leak may
be called noise. An effect that is a source of noise for one system may not be a
source of noise for another, depending on what measurements are made by the
system, the procedure by which they are made, and the analysis that is used to
derive leak rate information from these measurements.
The ability to detect a signal is limited by that portion of the noise energy with the
same frequency characteristics as the signal (i.e., that portion which could be
confused with the signal). The best way to characterize the noise field is to conduct
a large number of tests on one or more nonleaking pipelines over a wide range of
conditions. The statistical fluctuation of the noise is observed in the histogram* of
the volume-rate results created by plotting the measured volume rates from tests
conducted by a given system. The system's output when a leak is present, i.e., the
signal plus the noise, can be characterized by means of the relationship between the
signal and the noise. If it is not possible to determine what this relationship is, the
signal-plus-noise histogram must be measured for each leak rate at which one
wishes to know the performance of the system.
Throughout this document, the term "histogram" is used to mean "a graphical or numerical
representation of the likelihood that a quantity will be within a range of values." It is easily
derived from data and is the primary tool in evaluating system performance.
15
-------
An example of the histogram and the frequency distribution for a generic volumetric
leak detection system is shown in Figure 2.1. The frequency distribution describes
the fraction of the total number of test results in a defined interval. The likelihood of
exceeding a specified noise level is described by the integral of the frequency
distribution. The resulting cumulative frequency distribution is shown in Figure 2.2.
The cumulative frequency distribution is a more useful representation of the
histogram because it can be used directly in the performance calculations. If the
signal is constant over time and is independent and additive with the noise, the
signal-plus-noise histogram can be estimated directly from the noise histogram. For
this signal, the signal-plus-noise histogram has the same shape as the noise
histogram, but the mean of the noise histogram is equal to the signal strength. An
example of the cumulative frequency distribution of the signal-plus-noise histogram
for a leak of 0.10 gal/h (flowing out of the pipeline) is shown in Figure 2.3; this is for a
volumetric system. Statistical models of the noise and signal-plus-noise could also
be developed from the cumulative frequency distributions by means of standard
probability distributions, but no models are used in this protocol.
16
-------
-0.2
•O.2
-0,1 0
VOLUME RATE
-0.1 0 0.1
VOLUME RATE - gal*
Figure 2.1. Histogram (a) and frequency distribution (b) of the noise compiled from 25 leak
detection tests on nonleaking pipelines for a volumetric leak detection system. The mean
and standard deviation are -0.003 and 0.031 gal/h, respectively.
17
-------
1.0
o
D
•~^>
I i
to
o
SAU
i '
s
o
0 -
-0
-?
-Q.I
0.1
0.2
VOLUME RATE -
Figure 2.2. Cumulative frequency distribution of the noise derived from the frequency
distribution in Figure 2.1.
I
e?
10
tt
u,
yu
I
3
a
OS -
VOLUME RATE - gal/h
Figure 2.3. Cumulative frequency distribution of the signal-plus-noise generated for a leak
rate (i.e., signal) of-0.10 gal/h using the cumulative frequency distribution of the noise shown
in Figure 2.2.
18
-------
Figure 2.4 presents one statistical model, based on the cumulative frequency
distributions shown in Figures 2.2 and 2.3, that can be used to estimate the
performance of a detection system in terms of PD and PFA. The noise histogram,
represented by its cumulative frequency distribution and centered about zero, shows
the volume fluctuation level during tests in pipelines with no leaks. The dashed
curve reflects the cumulative frequency distribution of the signal-plus-noise
histogram from a pipeline with a leak of 0.10 gal/h. The model shown in Figure 2.4
can be used to determine the performance of the detection system against a 0.10-
gal/h leak; the performance against other leaks can be estimated by shifting the
signal-plus-noise cumulative frequency distribution accordingly. A leak is declared
whenever the measured volume rate exceeds the threshold. For a specified
detection threshold, T, the PFA is the fractional time that the noise will exceed the
threshold; the PFA is represented by the large dot on the cumulative frequency
distribution of the noise. In this example, the PFA equals 0.085. The PD is the
fractional time that the measured volume rate, with the signal present, will exceed
the threshold; the PD is represented by the large dot on the signal-plus-noise
cumulative frequency distribution. In this example, the PD equals 0.945. The
probability of a missed detection is 1.0 - PD.
o
I
o
UJ
DC
u.
UJ
2
O
-0,2
-0,1
0.1
0,2
-gal/h
Figure 2.4. Statistical model for calculating the PD and PFA of a pipeline leak detection
system. The model is set up to calculate the performance of the leak detection system
against a -0.10 gal/h leak rate. Fora threshold of -0.05 gal/h, the PFA = 0.085 and the PD =
0.945.
19
-------
The PD, PFA, T, and LR are all interrelated; changing one parameter affects the value
of one or more of the other parameters. The choice of parameters affects the
conclusions to be drawn from leak detection tests (i.e., the reliability of the test
result). Once the threshold has been selected, the PFA is determined and does not
change, regardless of the leak rate to be detected. The PD, however, does change
with leak rate if the threshold is kept constant. The PD increases as the detectable
leak rate increases, i.e., there is a better chance of finding large leaks than small
leaks. The threshold is usually chosen in such a way that the PD and PFA present an
acceptable balance between economic and environmental risks.
20
-------
SECTION 3
GENERAL FEATURES OF THE EVALUATION PROTOCOL
The protocol for conducting an evaluation consists of 13 basic steps. Before going into
these, however, we first examine how pipeline configuration impacts the evaluation and
how various options within this protocol can best be used.
3.1 PIPELINE CONFIGURATION
There is a wide range of pressurized pipeline systems that must be tested
periodically for leaks. The leak detection systems used in this kind of testing must
comply with the EPA regulation. The performance of many pipeline leak detectors,
especially pressure detection systems, will vary according to the configuration of the
pipeline system. The magnitude of the signal as well as that of the noise will be
affected. This occurs because the overall compressibility characteristics of the
pipeline system are influenced by the choice of material (fiberglass or steel), the use
of flexible hosing (and its length), and the presence of a mechanical line leak
detector* or other appurtenances. For example, the temperature- and leak-induced
pressure changes that occur in a static line are inversely proportional to the
compressibility of the pipeline system (see [4,5]). This interaction between the
pipeline and the performance of the leak detection system presents a challenging
problem: the same leak detection system can perform very well on one pipeline
system and poorly on another. Fortunately, the compressibility characteristics of the
line can be described by the bulk modulus, B, of the pipeline system, where B is the
inverse of K, the constant that describes the compressibility of the pipeline system.
Two pipelines may have different configurations, but may have the same
compressibility characteristics. In this protocol, B, which can be readily measured
(see Section 4.3), is used to characterize the pipeline used in the evaluation.
Pipelines constructed at special instrumented test facilities should simulate the
important features of the type of pipeline systems found at operational LIST facilities.
This protocol assumes that the leak detection systems to be evaluated are intended
for use on underground storage tanks that are typically 10,000 gal in capacity, where
the diameter of the pipe is typically 2 in. and the length is usually less than 200 ft. If
the leak the detection system will be used on pipelines with larger diameters or
A mechanical line leak detector is a device that has been used for many years at retail
petroleum stations to monitor the pipeline for the presence of large leaks. This device is
designed to detect leaks of 3 gal/h or larger defined at a line pressure of 10 psi. The hourly test
required by the EPA regulation is based on this device. Because of its wide use and its known
effect on the performance of pressure detection systems, it should be included as part of the
pipeline configuration if the leak detection system to be evaluated conducts a test while this
device is in the line.
21
-------
longer lengths, the evaluator should use a proportionately larger pipeline in
conducting the evaluation. Whether the evaluation is conducted at a special
instrumented testing facility or at one or more instrumented operational LIST
facilities, the minimum requirements are as follows.
• The pipeline, which can be constructed of either fiberglass or steel, must
have a diameter of at least 2 in. ± 0.5 in.
• The pipeline must be at least 75 ft long.
• The pipeline system must have a B of approximately 25,000 psi ± 10,000 psi.
• A mechanical line leak detector must be present within line if the leak
detection system being evaluated normally conducts a test with this device in
place.
• There must be a way to pressurize the pipeline system.
• There must be a tank or storage container to hold product withdrawn from the
line during a test.
• There must be a pump to circulate product from the storage container
through the pipeline for up to 1 h. (At operational LIST facilities and at most
test facilities, this container will be an underground storage tank, and a
submersible pump will be used to pressurize the pipeline and circulate
product through it.)
• The pipeline must have valves that can be used to isolate it from the storage
tank and the dispenser. These valves must be checked for tightness under
the maximum operating pressure of the pipeline system.
• The pipeline must contain a petroleum product, preferably gasoline, during
the evaluation.
• In addition, when an evaluation is done at a special test facility, there must be
a unit to heat or cool the product in the storage container.
When the evaluation is done at five or more operational LIST facilities that are
geographically separated, it will suffice if only one of the facilities meets these
criteria, with the exception of the bulk modulus criterion, which does not have to be
met by any of the facilities.
The performance of some of the systems that can be evaluated with this protocol will
decrease as the diameter and/or length of the pipeline increases. This is particularly
true for volumetric measurement systems that are directly affected by thermal
22
-------
expansion or contraction of the product in the pipeline. The performance estimate
generated by this protocol is considered valid if the volume of the product in the
pipeline system being tested is less than twice the volume of product in the pipeline
used in the evaluation. This is an arbitrary limitation because it does not take into
account the type of system, the method of temperature compensation, or the actual
performance of the system. It was selected to allow flexibility in the application of the
system. Thus, in selecting the length of the pipeline to be used in the evaluation one
should consider how the system will ultimately be used operationally. Because the
limitation is arbitrary, this protocol also allows the manufacturer to present a separate
written justification indicating why pipelines with capacities larger than twice the
capacity of the evaluation pipeline should be permitted. Concurrence with this
justification must be given by the evaluator. Both the written justification and
evaluator's concurrence must be attached to the evaluation report.
3.2 SUMMARY OF OPTIONS FOR ESTIMATING PERFORMANCE WITH THIS
PROTOCOL
To estimate the performance of a pipeline leak detection system, one must develop
histograms of the noise and the signal-plus-noise. Each histogram generated
according to this protocol requires a minimum of 25 independent tests. As shown in
Section 10.1, this number ensures that an estimate of the PD of 0.95 and the PFA of
0.05 can be made directly from the data and that the uncertainty in the estimate of
the PD and PFA as measured by the 95% confidence intervals, is approximately 5%.
This protocol provides five options for generating the data necessary to develop
noise and signal-plus-noise histograms. The first option is to conduct the evaluation
at an instrumented test facility specifically designed to evaluate pipeline leak
detection systems, and the second is to do it at one or more operational LIST
facilities that are specially instrumented to conduct the evaluation. Both of these
options require that the data be collected under a specific set of product temperature
conditions, which are measured as part of the test procedure, on a pipeline system
that has defined characteristics. The instrumentation is minimal and does not require
that temperature sensors be placed inside the pipeline. The next two options require
that data be collected over a period of 6 to 12 months, either at 5 operational LIST
facilities where the integrity of the pipeline systems has been verified, or at 10 or
more operational LIST facilities. The stations should be geographically located so as
to represent different climatic conditions. Each of the operational LIST facilities
selected should receive a delivery of product to the tank at least once per week.
Options 3 and 4 should provide approximately the same range of temperature
conditions specified in Options 1 and 2 because of seasonal variations in the
temperature of the ground and the temperature of the product delivered to the tank.
In the fifth option, a simulation is used to estimate the performance of the leak
detection system. This simulation is developed from experimentally validated
23
-------
mathematical models of all the sources of noise that affect the performance of a
particular system. These five options for developing a noise histogram are described
more fully in Section 6. It is assumed that the first four will be the most commonly
used; therefore, the last one is only briefly described.
3.2.1 Generating the Noise Histogram
The primary source of noise for a pipeline leak detection system is the thermal
expansion and contraction of the product in the line. Thus, the performance of
most pipeline leak detection systems is controlled primarily by temperature
changes in the product that is in the line. These changes are present unless no
product has been pumped through the pipeline for many hours. In order to take
these changes into account, the protocol described in this document requires that
all leak detection systems be evaluated under a wide range of temperature
conditions.
The range of temperature conditions used in this protocol is based on the results
of an analytical study of the climatic conditions found throughout the United
States [6,7]. The study estimated the average difference in temperature between
the product in the tank and the temperature of the ground around the pipe. The
results indicated that values of ±25°F would cover a wide range of conditions.
(This is the same range of temperature conditions generated for the EPA's
evaluation of volumetric leak detection systems [6,7].) All systems will be
evaluated in accordance with their own test protocols under a predetermined
matrix of temperature conditions created from an average of the product
deliveries and normal dispensing conditions throughout the United States. The
protocol in this document describes specifically how to create these conditions.
The performance of most detection systems is also affected by the pressure and
volume changes produced by the thermal expansion or contraction of any
trapped vapor in the line; in some instances, a leak detection device will simply
not work if vapor is trapped in the line. For this reason a significant effort should
be made to remove any trapped vapor. Trapped vapor will affect the
compressibility of the line and, thus, the magnitude of the bulk modulus. This
will, in turn, affect the magnitude of the calibration factor used to convert the
measured quantity (e.g., pressure changes) to volume changes. Even the
presence of small amounts of trapped vapor can be the source of large errors.
The presence of trapped vapor can be determined from the pressure-volume
data used to estimate the bulk modulus; vapor in the line should be suspected if
the pressure-volume curve is not linear but exhibits second-order curvature, as
illustrated in Figure 3.1, which shows the pressure-volume data obtained on a
200-ft, 2-in.-diameter pipeline at the UST Test Apparatus in Edison, New Jersey,
under two conditions: (a) with 105 ml of vapor in the line and (b) without any
24
-------
vapor in the line. Since the presence of trapped vapor can be easily checked
(see Sections 4.3 and 4.5), this protocol assumes that the leak detection system
being evaluated would test the line for vapor and either not test the line or would
remove it, if it is present, before a test is begun. As a consequence, all vapor
should be removed from the pipeline for all of the tests done and used in
estimating performance when the evaluation is conducted at an instrumented test
facility (i.e., Options 1, 2, and 5). To assess the sensitivity of the system to
trapped vapor, this protocol requires only a few tests to determine the sensitivity
of the leak detection system to vapor.
In this protocol, the data used to estimate the bulk modulus will determine
whether vapor is present in the line, and three special tests will be made with a
small volume of vapor trapped in the line to determine how the system performs
under this condition. The results of these three tests will not be included in the
performance estimates but will be presented in the evaluation report so that
manufacturer's claims about the effects of trapped vapor on the test results can
be better assessed.
A histogram of the noise is a requirement for making an estimate of the
probability of false alarm. The detection threshold is used to determine the
probability of false alarm directly from the histogram of the noise. The histogram
of the noise should be compiled from the results of pipeline leak detection tests
conducted over a wide range of environmental conditions and pipeline
configurations. The tests must be conducted on pipeline systems that are tight.
Temperature changes in the product in the line are the main source of noise
associated with the type of system likely to be evaluated with this protocol.
Therefore, a test matrix of temperature conditions has been defined. The
temperature conditions are based on those that might be encountered near the
end of the day at a moderate- to high-volume retail station.
25
-------
20
UJ
CC
Z3
to
t/>
UJ
cc
a.
10 -
» WITH VAPOR
* PIPELINE WITHOUT
I \^
^0
too
120
40 SB 80
- mi
Figure 3.1. Pressure-volume relationship for a 2-in.-diameter, 200-ft steel pipeline with
and without vapor trapped in the pipeline system.
3.2.2 Generating the Signal-plus-noise Histogram
A histogram of the signal-plus-noise is a requirement for making an estimate of
the probability of detection for each leak rate of interest. The threshold value is
used to determine the probability of detection directly from the histogram of the
signal-plus-noise for a given leak rate. A separate histogram of the signal-plus-
noise is required for each signal (i.e., leak rate) for which the performance in
terms of probability of detection is desired. For each leak rate of interest, the
histogram of the signal-plus-noise must be developed over the same temperature
conditions and pipeline configurations used to generate the noise histogram.
This protocol requires, at a minimum, that the probability of detection be
estimated against the leak rate specified in the EPA regulation for the type of
leak detection system being evaluated (i.e., 0.1, 0.2, or 3.0 gal/h). If a signal-
plus-noise histogram is developed for a second leak rate, an estimate of
performance can be made for a wider range of leak rates, because a relationship
between the signal and the noise can be developed.
Generating the signal-plus-noise histogram may be simple or may involve
significant effort. There are two options. The direct approach is to develop the
26
-------
histogram by generating a leak in the line and conducting a large number of leak
detection tests under the same conditions used to develop the histogram of the
noise. This direct approach can be used regardless of whether the leak
detection system uses a preset threshold or measures the flow rate directly.
Noise and signal-plus-noise histograms are required for each temperature
condition. In this approach, the histogram of the signal-plus-noise is measured
directly for the leak rate at which the probability of detection is desired, and thus
the relationship between signal and noise is determined directly. If the duration
of the leak detection test is short, the data necessary to develop the noise and
signal-plus-noise histograms can be acquired by conducting two tests in
succession. The direct approach is most beneficial when a PD is required for
only a few leak rates; otherwise, the time required to collect the data can be
excessive. This approach is easy to implement when data are collected at an
instrumented test facility or one or more instrumented operational LIST facilities,
but it is cumbersome if the data must be collected over an extended period at
many noninstrumented operational LIST facilities. If the probability of detection is
required for a large number of leak rates or if the test duration is sufficiently long
that only one leak detection test can be conducted for a given temperature
condition, the second approach would be more logical.
The second approach is to develop a signal-plus-noise histogram from the
histogram of the noise by developing a theoretical relationship between the
signal and the noise. An experimentally validated model that gives the
relationship between the signal and each source of noise must be developed.
With this model and the histogram of the noise, the signal-plus-noise histogram
can be developed for any leak rate, and an estimate of the probability of
detection can be made for any leak rate. This relationship must be valid over the
range of test conditions and pipeline configurations covered by the evaluation. It
can be used with all five of the options for data collection. It is particularly useful
for evaluating the performance of leak detection systems that require long tests
or long waiting periods or that acquire the noise data at many operational LIST
facilities over a long period of time.
Developing the relationship between the signal and the noise can be difficult if
these two phenomena are coupled (i.e., if the noise affects the magnitude of the
signal). This occurs, for example, if the pressure, volume or flow-rate changes
produced by a leak do not add in a one-to-one manner with the pressure, volume
or flow-rate changes produced by each noise source (e.g., temperature changes
of the product in the pipeline). If the signal does add linearly with the noise, such
a relationship is easily developed by fitting a curve to a plot of the output of the
measurement system versus the actual leak rate for two or more leaks generated
under benign noise conditions. This curve gives the relationship between the
27
-------
output of the measurement system and the flow rate due to a leak. If the leak
detection system is one that measures volume, developing the relationship
between the signal and the noise is relatively straightforward, because the
volume changes produced by thermal expansion or contraction usually add to
those produced by a leak. If, however, the system is one that measures
pressure, developing this relationship is more difficult, especially when thermal
changes in the product are not compensated for. Not only do the measurements
have to be converted from units of pressure to units of volume, but the
relationship between pressure and volume is not constant; it changes with
pipeline configuration and may also change as a function of the time elapsed
since the last change of pressure in the pipeline.
A detailed explanation of how to develop the relationship between the signal and
the noise will not be presented here; there are many ways to develop the
relationship and many to verify that the relationship is correct. It is up to the
manufacturer of the leak detection system to do this. This protocol requires that
the relationship be verified with a simple measurement procedure, which is
described in Section 4.2.3. This procedure should be undertaken before the
noise data are collected. If the relationship has not been verified, the signal-plus-
noise histogram must be developed directly during the evaluation procedure.
3.2.3 Generating Histograms with Leak Detection Systems that Use a
Multiple-test Strategy
There are many possible schemes for implementing a multiple-test strategy. A
leak may be declared if the threshold is exceeded in a certain number of test
sequences, for example, one out of two, two out of two, or two out of three test
sequences, any other m-out-of-n scheme, or the average of two or more tests.
These are only a few examples. The most common multiple-test strategy is to
conduct a second test only if the threshold is exceeded in the first test. The
critical factor is that the data used to build the histograms must come from that
test sequence which was the basis for declaring a leak. For example, when a
second test is conducted only if the threshold is exceeded during the first test,
this means the last test in the sequence; if the threshold is not exceeded the first
and last tests are by definition the same. When two or more tests are always
required, this means the smallest test result out of the n tests conducted or the
average of all of the tests. In addition to histograms used to develop a
performance estimate of the system, a second performance estimate is
requested. This second estimate is based only on the results of the first test in
the multiple-test sequence.
28
-------
3.3 CONDUCTING THE EVALUATION
The protocol, which is summarized below, requires that a leak detection system be
evaluated under a wide range of pipeline configurations and test conditions. It can
be used to evaluate systems that require multiple tests as well as those based on a
single test.
Step 1 - Describe the leak detection system. The first step in an evaluation is to
specify the important features of the leak detection system. This step is important for
three reasons. First, a brief description will identify the system as the one that was
evaluated. Second, changes to the system may be made at a later date, but the
manufacturer may not feel that the changes are important enough for him to rename
the system. Such changes may affect the performance, either for better or worse. If
the characteristics of the system have been specified in a brief descriptive statement,
the owner/operator of an underground storage tank system will have a way to
determine whether the detection system he is using is actually the one that was
evaluated. Third, the owner/operator will be able to interpret the results of the
evaluation more easily if he has this information.
The description of the leak detection system need not be excessively detailed, and
proprietary information about the system is not required. The description should,
however, include the important features of the instrumentation, the test protocol, and
detection criterion. If the system requires multiple tests before a leak is declared,
this should be clearly stated. A summary sheet on which to describe the system is
provided as Attachment 1 in Appendix B.
Step 2 - Select an evaluation option. The second step is to determine which one
of the five evaluation options will be used: test facility, one or more instrumented
operational LIST facilities, 6- to 12-month data collection effort at 5 operational LIST
facilities at which pipeline integrity has been verified, 6- to 12-month data collection
effort at 10 or more operational LIST facilities, or validated computer simulation.
Step 3 - Select temperature and leak conditions for evaluation. The third step is
to define the temperature and leak conditions under which the evaluation will be
performed. If the evaluation is done at a test facility, at one or more instrumented
operational LIST facilities, or by computer simulation, the temperature conditions
necessary to compile the noise histogram will be developed according to a test
matrix, which is generated before the data collection begins, and verified by means
of specific diagnostic ground and product measurements made immediately before
the test. A matrix of leak conditions will also be generated so that a histogram of the
signal-plus-noise can be compiled; the type of test matrix will depend on whether the
leak rates are known a priori or whether a blind-testing procedure is used.
29
-------
If the data are collected at operational LIST facilities over a period of 6 to 12 months,
temperature conditions do not need to be artificially generated, but the relationship
between the measured quantity and the flow rate that would be produced by a leak
at the manufacturer's standard test pressure (i.e., the relationship between the signal
and the noise) should be defined and provided by the manufacturer before an
evaluation of the system is performed. This relationship is used to generate the
signal-plus-noise histogram from the noise histogram at the EPA-specified leak rate.
The relationship can be either a theoretical one that has been validated
experimentally or an empirical one that has been developed through
experimentation.
Step 4 - Assemble equipment and diagnostic instrumentation. The fourth step
is to assemble the equipment needed for the evaluation and to calibrate diagnostic
instrumentation such as pressure and temperature sensors.
Step 5 - Verify the integrity of the pipeline system. Conducting a performance
evaluation of a leak detection system requires a nonleaking pipeline. If the pipeline
is not tight, the performance of the system being evaluated will be degraded. For all
but one of the evaluation options (Option 4) presented in this protocol, it is
recommended, though not required, that the integrity of the pipeline be verified
beforehand with a leak detection system whose performance is already known.
Step 6 - Determine the characteristics of the pipeline system. The sixth step is
to determine whether the pipeline system used in the evaluation meets the minimum
specified conditions. The same pipeline configuration can be used regardless of
whether the evaluation is done at a test facility, one or more instrumented
operational LIST facilities, or by the simulation approach. The compressibility of the
pipeline system must be within a specified range; if it is not, a mechanical device can
be used to modify the compressibility characteristics of the line for the test. An
example of a device that can be used to modify the compressibility characteristics of
the pipeline system is described in Section 4.3.
Step 7 - Evaluate-the performance characteristics of the sensor subsystems.
The seventh step is to characterize the performance of the measurement
subsystems (instrumentation). The resolution, precision, accuracy, minimum
detectable quantity, and what the instrumentation is measuring (i.e., specificity) must
be determined. Also, the flow rate at the threshold must be determined. Although
this step is not actually required in order to estimate the performance of the system,
it serves two important purposes. First, it indicates, before the evaluation is
performed, whether the instrumentation is working according to the manufacturer's
specifications. If the instrumentation is not performing properly or if it is out of
calibration, the evaluation should not proceed until the problems are remedied.
Second, the instrumentation will ultimately limit the performance of the leak detection
30
-------
system. If it is evident that the performance expectations of the manufacturer are
more than what the instruments will allow, the evaluation can be stopped before too
much time has been invested or too much expense incurred. Furthermore, this step
can be completed quickly.
Step 8 - Develop (if necessary) a relationship between the leak and the output
of the measurement system. If the relationship between the leak and the output of
the measurement system (i.e., between the signal and the noise) is known or has
been supplied by the manufacturer and no direct estimate of the signal-plus-noise
histogram at the EPA-specified leak rate has been made as part of this protocol,
experiments must be conducted to verify the relationship. This step is not necessary
if the test matrix requires the conduct of 25 tests at the EPA-specified leak rate (i.e.,
developing the signal-plus-noise histogram with the direct approach).
Step 9 - Develop a histogram of the noise. The ninth step is to develop a
histogram of the noise under the temperature conditions specified in Step 3 for the
pipeline system specified in Step 6. This histogram, which is needed to estimate the
probability of false alarm, is generated from one or more pipeline tests, conducted
according to the manufacturer's protocol, for each condition given in Step 3. If the
system uses a multiple-test procedure, two histograms are required. The
performance of the system, which includes the entire multiple-test sequence, is
generated from the data from the test result used to determine whether the pipeline
is leaking (in many instances these are the data from the last test in the sequence).
Step 9 is the heart of any evaluation. Once the histogram of the noise is known and
either the relationship between the signal and the noise is known or a histogram of
the signal-plus-noise has been developed, the performance of the system can be
estimated.
Step 10 - Develop a histogram of the signal-plus-noise. The tenth step is to
develop a histogram of the signal-plus-noise for each leak rate at which the system
will be evaluated and under the same conditions used to generate the noise
histogram. If system uses a multiple-test procedure, two histograms are required.
The performance of the system, which includes the entire multiple-test sequence, is
generated from the data from the test result used to determine whether the pipeline
is leaking (in many instances these are the data from the last test in the sequence).
This histogram is needed to estimate the probability of detection.
It may be a simple matter to generate the histogram, or it may involve significant
effort. The histogram of the signal-plus-noise may be measured directly for each
leak rate of interest by developing a histogram of the test results when a leak of a
given magnitude is present. As an alternative, a model may be developed and
validated experimentally that gives the relationship between the signal and the noise.
As stated in Section 2.3.2, if the relationship between the signal and noise is known,
31
-------
the noise histogram can be used to estimate the signal-plus noise histogram. This
relationship can be difficult to develop unless all sources of noise during the test are
compensated for (or unless they are small). A model is required if one wants to
know a system's performance at many leak rates that are different from those
specified in the EPA regulation.
Step 11 - Determine the system's sensitivity to trapped vapor. The eleventh
step is to determine the sensitivity of the leak detection system to vapor trapped in
the pipeline system. To this end, three special leak detection tests will be performed.
Step 12 - Conduct the performance analysis. The twelfth step is to calculate the
performance of the system in terms of PFA and PD at the EPA-specified leak rate.
The protocol is designed so that the PD and PFA of the system are determined with
the manufacturer's threshold at the leak rate and test pressure specified by the EPA
regulation (i.e., 0.1, 0.2, or 3 gal/h). If the evaluation is not done at the pressure
specified by the EPA, a method is given to calculate an equivalent leak rate at
whatever pressure is used. The protocol provides, as Attachment 2 in Appendix B, a
summary sheet to be used in reporting a variety of other performance estimates so
that the performance can be compared to that of other leak detection systems. If a
system uses a multiple-test procedure, the protocol requires a second performance
estimate based on noise and signal-plus-noise data from the first test of the multiple-
test sequence.
Step 13 - Evaluation report. The thirteenth and final step is to report the results of
the evaluation in a standard format, given in Appendix A. This form has seven
attachments, which are provided in Appendix B. The performance characteristics of
the instrumentation, the estimates of the system's performance in detecting leaks in
the ambient environment, and the sensitivity of the system to trapped vapor will be
presented in a set of tables. The test conditions and pipeline systems to which the
detector is applicable will also be presented.
3.4 ACCURACY OF THE EVALUATION
The accuracy of the evaluation basically depends on whether the noise and signal-
plus-noise histograms were generated under the required range of temperature
conditions, whether the test result was influenced by the fact that the flow rate from
the pipeline was known, and whether one or more test results was removed from the
data set without adequate justification. In general, a performance estimate will tend
to be unrealistically optimistic if (1) less than the full range of temperature conditions
was used in the evaluation, (2) part of the test protocol was changed, such as the
duration of a waiting period or the duration of the actual test, or (3) one or more of
the test results was removed arbitrarily. In the first case, because the temperature
matrix consists of a range of conditions, the index used to characterize the
32
-------
temperature conditions has an uncertainty associated with it; contributing to the
second error is the fact that many of the test protocols for the leak detection systems
are not definitive enough or require some intervention on the part of the operator,
whose judgment can be influenced if he knows the status of the pipeline during a
test; in the third case, an anomalously large test result might be removed simply
because it did not match the expected leak rates. Accurate evaluations can best be
assured
• by carefully following the evaluation protocol
• by defining the leak detection protocol before the evaluation begins and following
it carefully throughout the evaluation
• by using all of the data collected during the evaluation in the performance
analysis
The use of, or the failure to use, all the data tends to have the most significant impact
on the results of an evaluation. Estimates of the probability of false alarm and the
probability of detection are made from the test results that comprise the tails of the
noise and signal-plus-noise histograms. When only 25 tests are used, an estimate of
a probability of detection of 0.95 or an estimate of a probability of false alarm of 0.05
depends on only one or two test results. Improperly removing one of these from the
data set can significantly alter the performance estimates. Therefore, once an
evaluation is begun, all of the data should be used unless the leak detection system
or the equipment at the evaluation facility can be shown to be malfunctioning, or the
evaluation procedure is not being properly implemented. If test results are removed
from the data set used to generate either the noise or the signal-plus-noise
histogram, this must be clearly indicated, explained, and justified in the evaluation
report.
The evaluator (either the manufacturer or a third party) has the option of developing
histograms of the noise and the signal-plus-noise with full knowledge of the leak rate
in the pipeline during a test. Or, he may opt for a blind testing procedure, which in
practice requires that the evaluation be done at a test facility or one or more
instrumented operational LIST facilities. In a full-scale blind test, the actual flow rates
and temperature conditions would not be made available to the test crew until the
entire evaluation had been completed. With the protocol used here, however, the
test crew knows that one of the leak rates will be zero and one will be the EPA-
specified leak rate (i.e., 0.1, 0.2, or 3.0 gal/h). The only possibility, then, is a partially
blind test, in which the order of the leak rates is unknown or in which a small
percentage of the leak rates is different from the EPA-specified leak rate, or both.
One of the partially blind testing procedures used in this protocol requires that 10 to
20% of the leak rates be changed without the knowledge of the test crew. If any of
33
-------
these larger test results is arbitrarily removed, the evaluation is declared invalid and
must be repeated. (Temperature conditions can be manipulated in the same way as
leak rates.) The partially blind test is intended for use by a third party evaluator, but
can also be used by a manufacturer.
3.5 OTHER ACCEPTABLE EVALUATION PROTOCOLS
This evaluation protocol is designed to cover most leak detection systems that
measure pressure changes or losses in the volume of product in the pipeline. It is
consistent with the ASTM practice [8] being developed for evaluating and reporting
the perfomlance of leak detection devices used on LIST pipeline systems. There
may exist leak detection systems to which this protocol cannot be easily applied, or
there may be additional variations of this protocol that might be easier to implement.
Other methods of evaluating performance which follow the general approach in
Section 2.2 are also acceptable providing that the test conditions are at least as
stringent as those described here and that the required number of pipeline
configurations is at least as great. Alternative methods of evaluation, which are
acceptable to the EPA, are presented in the Preface of this document.
34
-------
SECTION 4
EQUIPMENT NEEDED FOR GENERATING
EVALUATION CONDITIONS
The conditions that one must be able to generate or modify during an evaluation are:
line pressure, which influences the leak rate; the leak itself; the compressibility of the
line; the temperature of the product in the line; and the amount of vapor trapped in the
line. Depending on which of the five evaluation options is selected, one or more pieces
of equipment may be required: a leakmaker, a mechanical device to modify the
compressibility of the pipeline system, a mechanical device to trap vapor in the pipeline
system, a pressure sensor, and tank and ground temperature sensors. This equipment
should meet the following guidelines:
• It should measure the flow rate due to a leak in the line at a specified pressure
with an accuracy of 0.01 gal/h.
• It should measure the bulk modulus, B, of the pipeline system with a precision
and accuracy such that B/V0 is known within 0.025 psi/ml, where V0 is the volume
of the product in the pipeline.
• It should measure the total volume of product in the line to within 1 gal.
• It should measure the difference in temperature between the ground and the
product at the bottom of the tank (which is brought into the pipeline to produce a
temperature condition) with an accuracy of 0.2°F.
• It should measure line pressure during the test with a precision of 0.5 psi and an
accuracy of 1 psi or better.
This protocol recommends certain equipment and procedures for making these
measurements but does not limit the choice of equipment or procedures to these alone.
The protocol requires only that the measurements be made within the specified range of
precision and accuracy, and under the specified range of conditions.
4.1 LINE PRESSURE
A pressure sensor is necessary to determine the pressure in the line during each test
and to set a leak rate. Pressure measurements can be made with either a
mechanical gauge or an electromechanical transducer and automatic data
acquisition system. A mechanical gauge that has been calibrated is more than
satisfactory.
35
-------
4.1.1 Equipment and Instrumentation for Generating Line Pressure
A mechanical pressure gauge that can be read manually to the nearest 0.5 psi
and has an accuracy of 1 psi can be used to measure pressure. To measure
pressure automatically, a pressure transducer that has a precision and accuracy
of 0.5 and 1 psi, respectively, can be used. Even if pressure is recorded
automatically, it is desirable to insert a mechanical pressure gauge in the line to
help conduct and control the experimental measurements. The pressure sensor
can be attached at any point on the pipeline.
These pressure sensors should be calibrated before each evaluation, or more
frequently, if required. Calibration is done by applying a known pressure to the
system and recording the output of the sensor. A mercury manometer can be
used for this purpose. Calibration data should be obtained in increments of 5 psi
or less. At least five points are required. A calibration curve is generated by
fitting a regression line to the pressure measured by the sensor being calibrated
(y axis) and the known pressure from the reference source (x axis). The
precision of the sensor is estimated from the standard deviation of the ordinate (y
axis). The accuracy is determined from the intercept of the curve of the leak rate.
The calibration curve should be used to convert the output of the sensor to
pressure units (e.g., volts to psi); if the sensor output is already in units of
pressure, the calibration curve will correct any measurement errors that the
sensor may have developed since its original calibration by the manufacturer.
4.1.2 Measurement of Line Pressure
If pressure measurements are recorded digitally by a computer, it is important
that the time clocks on all the instruments be synchronized to the nearest second
with the clock used in the evaluation, and that the start and end times of all
pressure measurements required to complete the evaluation be recorded. If the
pressure measurements are made with a mechanical or electrical gauge, the
pressures should be read by the tester and the time of the reading recorded.
4.2 LEAK RATE
One or more leaks must be generated during an evaluation as a means of
developing a signal-plus-noise histogram. A device is needed that can establish and
maintain a leak with a constant flow rate at a given pressure. This can be done, for
example, by using a flow meter or by measuring the volume of product that is
released over time through a valve or orifice. This protocol shows how a leak can be
generated with the latter approach (see Sections 4.2.1 and 4.2.2), but any device will
do provided that it is properly calibrated and used. For example, if a flow meter set
to generate a particular flow rate is used, the flow rate must be verified
36
-------
experimentally at the appropriate pressures by means of a method similar to the one
described in Section 4.2.2.
A leak can be generated at any location in the line. Generally, it is most convenient
to withdraw product at either end of the line, i.e., either near the submersible pump
and mechanical line leak detector or at the shear valve near the dispenser. The
latter tends to be the easiest location at which to generate and measure the leak.
This protocol has established a line pressure of 20 psi as the standard pressure for
defining a leak rate for all pipeline leak detection systems, with the exception of the
hourly testing systems, in which the EPA regulation has established a specific
pressure of 10 psi (i.e., 3 gal/h) as the standard for defining the leak to be detected.
As a consequence, all values of leak rate will he established at 10 psi for the hourly
testing systems designed to meet the 3-gal/h EPA standard and at 20 psi for all other
systems designed to meet the 0.2-gal/h monthly monitoring or 0.1-gal/h line tightness
testing EPA-standards. When using a leak-making device similar to the one
described in Section 4.2.1, the evaluator sets a leak rate by adjusting the size of an
orifice, usually by means of an adjustable valve. Once the rate of the leak through
the valve or orifice has been set at either 10 psi or 20 psi, depending on whether the
system uses an hourly test or not, any other pressure can be used during the
evaluation provided that the size of the orifice does not change. For any system
being evaluated, an initial test pressure will be stipulated by the manufacturer; it is
recommended that the leak rate be measured at this initial pressure in addition to the
10 or 20 psi.
If it is not possible to establish the leak rate at 10 or 20 psi, the appropriate leak rate
for the given pressure can be established by means of a mathematical relationship.
This mathematical relationship can be used to determine the equivalent leak rate at
the test pressure so that the EPA-specified leak rate is properly defined at 10 or 20
psi.
The mathematical relationship required to convert a leak rate generated at the test
pressure to 20 psi depends on whether the flow is laminar or turbulent, which in turn
depends on the density and viscosity of the product, the diameter of the hole, and
the length and roughness characteristics of the leak-making apparatus itself. The
relationship describing the flow through a hole in an in situ pipeline is even more
complicated because the surrounding backfill and any residual sediment in the
product will also affect the flow rate. For laminar flow, the flow rate for free flow
through an orifice is proportional to the pressure at the orifice; for turbulent flow, the
flow rate is proportional to the square root of pressure. Eqs. (4.1) and (4.2) give
relationships that can be used to convert the leak rate at the test pressure to the leak
rate at 20 psi for turbulent and laminar flow, respectively. These equations can be
used to convert leak rate, LR, measured in psi at one pressure, P, to a leak rate,
37
-------
LR2oPsi, at a pressure of 20 psi. These two equations should bracket the actual
relationship for the pipeline, leak-maker and product.
(4.1)
(4.2)
This mathematical relationship should be developed empirically for the pipeline,
product and leak-making device to be used in the evaluation. This can be done by
setting the leak rate of interest at 10 or 20 psi and then measuring the same flow rate
through the same orifice at the test pressure; this procedure should be repeated
three times to obtain a median value. Once this has been done, the leak rate
measured at the test pressure can be used during the evaluation. It is important to
note that this leak rate will be different from but equivalent to the leak rate measured
at 10 or 20 psi.
Sometimes it is not possible to develop an empirical relationship. In such cases a
theoretical relationship can be used. If it is not possible to justify experimentally the
use of either Eqs. (4.1) or (4.2), Eq. (4.1) should be used. For gasoline motor fuels,
Eq. (4.1) agrees well with experimental measurements.
4.2.1 Equipment and Instrumentation for Generating Leaks
To generate the leak described above, the following equipment can be used: a
leak-making device that allows a constant flow of product from a pipeline,
graduated cylinders, a stopwatch, a pressure sensor, and a 1-gal storage
container that can safely handle petroleum fuels. Figure 4.1 illustrates the
important features of an apparatus that can be used to generate a leak. A
mechanical system that has three valves and that can be easily attached to and
detached from the line is required. One of the valves (Valve B) is a metered
valve that is used to set the leak rate and release product from the line. This
valve should have a dial mechanism that can be used to adjust and maintain a
constant flow rate. Another valve (Valve A), located between the line and the
metered valve, is used to open and close the line. Valve C is used to release a
larger volume of product from the line. One generates a leak at a given line
pressure by first pressurizing the line, then opening Valve A and adjusting Valve
B until the desired leak rate is obtained.
38
-------
VERNIER
THREADED f~~1
CONNECTION VALVE A ^ 1 J
TO
VALVE B
VALVE C
Figure 4.1. Schematic diagram of an apparatus to generate small and large leaks in the
pipeline.
4.2.2 Measurement of Leak Rate
The line must be kept at a constant pressure while the leak rate is being
measured. Normally, this would be the operating pressure of the pipeline during
dispensing of product.
Making this measurement requires a number of graduated cylinders, preferably
10 ml, 25 ml, 100 ml, and 250 ml in size. It is recommended that at least one
graduated cylinder of each size be available. Note that these cylinders should
not be used to store product; for safety reasons, a proper storage container
should be used to hold product removed from the pipeline during the tests.
The procedure for generating a leak is as follows:
• Bring the line to the pressure required for testing.
• Open Valve A and adjust Valve B until the leak rate of interest is
obtained. Then close Valve A until it is time to generate a leak in the line.
Open Valve A to generate the leak.
• Using a graduated cylinder and a stopwatch, measure the volume of
product released from the line until Valve A is closed. Recommendations
for the size of the graduated cylinder and the approximate lengths of the
measurements, in seconds, are given in Table 4.1. In general, these
measurements will be made in milliliters and will have to be convened to
gallons.
• Repeat the leak rate measurement twice and use the median of the three
leak rate estimates if the difference between the minimum and maximum
values is less than 0.02 gal/h.
39
-------
• Make additional measurements if the difference between the minimum
and maximum values exceed 0.02 gal/h, and use only the last three
consecutive measurements to make the calculation.
• Keep the pressure constant to within ±1 psi during the measurements.
Table 4.1. Recommendations for Measuring Leak Rate
Graduated
Cylinder Size
(ml)
10
10
25
25
100
100
Minimum
Graduated
Divisions*
(ml)
0.2
0.2
0.2
0.2
1.0
1.0
Length of
Measurement**
(s)
10
60
10
60
10
60
Leak Rate
(gal/h)
0.95
0.16
2.37
0.40
9.50
1.58
Error in
Measuring
Leak Rate
(gal/h)
0.009
0.002
0.009
0.002
0.047
0.008
* Read the graduated cylinder to the nearest 0.5 division.
** Record time to the nearest 0.1 s.
The leak rate should be measured each time the metered valve (Valve B) is
adjusted. The leak rate should also be checked if testing is done over a period of
1 h or longer at one set leak rate. When the test is long, it is recommended that
leak rate measurements be made at the beginning and end of the test period and
that the average leak rate be reported.
It is recommended that a calibration curve be developed for the metered valve so
that the dial on this valve can be used to set the approximate leak rate. This
calibration curve is generated at a specific pressure; five leak rates are
generated over the range of interest. A calibration curve can be developed by
fitting a regression line to an x-y plot of the dial readings (y axis) versus the
measured leak rates (x axis). This curve can be used to help control and simplify
the experimental procedure because it allows the evaluator to set the leak rate.
The dial should not be used to set leak rates unless the leak-generating
apparatus can be shown to have highly repeatable results, i.e., within 0.01 gal/h.
4.2.3 Relationship Between the Signal and the Noise
If the signal-plus-noise histogram required by this protocol at the EPA-specified
leak rate will be developed directly from measurements made during the
evaluation, it is not necessary to identify the relationship between the signal and
the noise, and the reader can proceed to Section 4.3.
40
-------
There are many approaches that can be used to verify that the relationship
between the signal and the noise provided by the manufacturer is valid. A
complete experimental validation requires that the histogram of the signal-plus-
noise be developed for at least three leak rates over a wide range of noise (i.e.,
temperature) conditions. This, however, constitutes more data than what is
obtained by directly measuring the signal-plus-noise histogram at the EPA-
specified leak rate. The amount of data that are necessary can be reduced
somewhat if the relationship between the signal and noise is based on well-
known physical models whose important features can be verified. If the
relationship is incorrectly defined, the performance of the leak detection system
could be adversely affected; the direct measurement approach, on the other
hand will not be impacted by an incorrect relationship. It is recommended,
therefore, that if the relationship between the signal and the noise has not been
thoroughly validated before the evaluation, it should not be used, and the signal-
plus-noise histogram should be generated from direct measurements.
This protocol requires two simple checks whose purpose is to determine whether
the relationship provided by the manufacturer is valid and can be used to develop
a signal-plus-noise histogram from the noise data.
The first check determines whether or not the relationship can be used to find the
mean of the signal-plus-noise histogram for a given leak rate. It also gives the
relationship between the output quantity and leak rate. Leak rates of
approximately 0.0, 0.05, 0.10, 0.20, 0.30, and 0.40 gal/h should be used if the
system is designed to detect either a 0.1- or a 0.2-gal/h leak rate, and rates of
approximately 0.0, 2.0, 2.5, 3.0, 3.5, and 4.0 gal/h should be used if the system is
designed to detect leaks of 3 gal/h. The leak should be generated at a constant
pressure of 10 or 20 psi, whichever is appropriate. If this is not possible, a leak
rate equivalent to the one specified at 10 or 20 psi can be generated at a
constant pressure other than 10 or 20 psi. An x-y plot of the output quantity of
the system (y axis) and the actual flow rate due to each (x axis) should be made,
and a regression line (least-squares line) should be fit to the data. The equation
that describes this line gives the relationship between the measured and actual
signal when the temperature changes are small. The output of the measurement
system calculated from this regression line (at the EPA-specified leak rate)
should then be compared to the output derived from the relationship provided by
the manufacturer. The standard deviation of the ordinate (y axis), an indication
of the uncertainty of the relationship, should also be calculated.
If there were a way of knowing a priori whether the signal adds linearly to the
noise, this check would be the only one required. Since it is not possible to know
this beforehand, both checks must be done. The first check does not assess
whether the relationship correctly predicts how the effects of a leak and product
41
-------
temperature changes are combined. If the signal does not add linearly with the
noise, the shape of the noise histogram (which might be assessed from the
standard deviation of the data) will be different from the shape of the signal-plus-
noise histogram, and additional information is required to check the relationship.
The second check verifies the relationship in cases when the temperature
changes in the product in a leaking pipeline are not small. It is this step that
could require significant effort. In this protocol, however, only a simple check is
done; if the manufacturer's relationship is verified by this check, it is assumed
that it is valid in general. Three leak detection tests are conducted according to
the procedures for generating a temperature condition in Section 5.1. The first
test (Test A) is done on a pipeline in which the temperature changes are
negligible. A leak equal to the EPA-specified leak rate is generated for this test.
The other two tests are done when there is at least a 10°F temperature
difference between the product in the pipeline and the temperature of the backfill
and soil surrounding the pipeline (these changes should be the same for each
test); one of the tests (Test B) is done on a tight pipeline and the other (Test C) is
done on a pipeline with a leak equal to the EPA-specified leak rate. When the
outputs from Tests A and B are combined according to the relationship provided
by the manufacturer, they should be equal to the output from Test C. That is, the
leak rate under the given temperature condition should equal the sum of (1) the
leak rate when there is no temperature change and (2) a zero leak rate under the
given temperature condition when (1) and (2) are properly combined.
There are no specified criteria in this protocol for accepting or rejecting either
check. The checks are made and the results are reported. If the checks show
that using this relationship will result in a large error, the relationship should not
be used. (Errors equivalent to 0.03 to 0.06 gal/h can have a significant impact on
the performance of the system against leak rates of 0.1 and 0.2 gal/h,
respectively.) The decision to use the relationship is up to the manufacturer.
The results of these tests should be reported in the tables provided in
Attachments 3, 4, and 7 in Appendix B. Attachment 7 summarizes the results of
the two checks. Attachments 3 and 4 summarize the temperature and leak
conditions, as well as the test results.
4.3 PIPELINE COMPRESSIBILITY CHARACTERISTICS
In four of the five evaluation options, the compressibility characteristics of the
pipeline system used in the evaluation must be determined. For three of these
options, this protocol gives a specific value for the compressibility of the line. The
compressibility can be characterized by the bulk modulus, B, of the pipeline system,
which can be estimated with a simple measurement procedure.
42
-------
4.3.1 Equipment and Instrumentation for Modifying Pipeline
Compressibility
To determine the compressibility of the pipeline, one needs a pressure sensor,
either mechanical or electrical, a leak-generating apparatus, a stopwatch, and a
graduated cylinder. If the compressibility characteristics of the pipeline do not
meet the specifications of the protocol (i.e., 25,000 psi ± 10,000 psi), there are
two choices: use another pipeline system or modify the compressibility
characteristics of the pipeline using the device shown in Figure 4.2.
TO
TO
t
fy~)
SPRINGS
IWW\
BLEED
oJoL
AIR
PRODUCT
Figure 4.2. Mechanical device to modify the compressibility characteristics of the
pipeline system.
The device shown in Figure 4.2 consists of a liquid-tight piston that is installed in
a cylinder. Liquid from the pipeline is allowed to enter the chamber in front of the
piston. When the pipeline is placed under pressure, the liquid will apply a force
on the face of the piston; the springs attached to the back of the piston resist this
force. This device will affect the compressibility of the pipeline system. The
magnitude of its effect depends on the spring constant.
A device of this type was built and used to modify the compressibility
characteristics of the pipeline system at the EPA's LIST Test Apparatus. The
43
-------
device consisted of a pneumatic cylinder* 2 in. in diameter and 12 in. long, a
piston with a stroke of 8 in., and two springs each having an outer diameter of-
11/16 in., a length of 4 3/4 in., and a spring constant of 11.9 psi. The device
changed the compressibility characteristics of the pipeline by a factor of three.
4.3.2 Measurement of Pipeline Compressibility
The procedure for measuring pipeline compressibility is to drain product from a
line initially raised to operating pressure, and then measure simultaneously the
cumulative volume of product released from the line and the pressure in the line
at the time of the volume measurement. The procedure includes the
compressibility effects of any vapor trapped in the line. If no vapor is trapped in
the pipeline, pressure (y axis) should be linearly related to volume (x axis). The
slope of a regression line fit to these data gives an estimate of BV0; B can be
estimated directly if the volume of the product in the line, V0, is known. Figure
4.3 is an example of the pressure-volume plot for data collected on a 2-in.-
diameter, 165-ft-long pipeline with and without a mechanical line leak detector
present. Figure 4.3 shows that the pressure-volume relationship is linear and
that it changes if a mechanical line leak detector is present.
The device that was assembled and field-tested during the development of this protocol was
built with a Chicago Pneumatic Cylinder Model DS-96-8-V.
44
-------
tu
re.
to
UJ
cc
CL.
a
Ul
CC
i
UJ
re
5 "
Figure 4.3. Pressure-volume relationship for a 2-in.-diameter, 165-ft pipeline (a) without
and (b) with a mechanical line leak detector.
Figure 4.4 shows the difference in the pressure-volume relationship in a 2-in.-
diameter, 200-ft-long steel pipeline when the compressibility device is attached to
the line and when it is not. If vapor is trapped in the pipeline, the pressure-
volume relationship will not be linear but will exhibit curvature as illustrated in
Figure 3.1. Thus, this measurement also provides a simple way to determine if
there is any vapor trapped in the pipeline.
45
-------
fcW -"
«™
-
~
75 ~
Q.
Ul
§ 10-
to
to
uu
cc
a.
-
™
-
0-
t
o
*«,, +
* Oo0 « PIPELINE WITH COMPRESSIBILITY DEVICE
t * °*0
+ ^Og
* °°«'
**«
+ **.
°«-0
*
,
* *«•«
°*»,
t *o0
* ''o
4-
, , i • ' i -i " r i ~i " rf- • f — — T — -f
o ao « to to 100 ta
- nil
Figure 4.4. Pressure-volume relationship for a 2-in.-diameter, 200-ft steel pipeline when
the compressibility device is attached to the line and when it is not.
The value of B will depend on when and how the test pressure in the line is
established. If the pressure is raised or lowered suddenly, as typically happens
when the submersible pump is turned on, the pressure changes in the line will be
adiabatic. If a test is conducted immediately after the pressure has been raised
suddenly and if the duration of the test is short (less than 5 min or so), B will be
nearly adiabatic. If the test is long (about 1 h) or if the pressure is kept constant
for 15 min before beginning a test, B will not be adiabatic and will have a different
value.
The pressure measurements are best accomplished with a mechanical pressure
gauge, which eliminates the time registration problems that are encountered if
volume measurements are made manually and if pressure measurements are
made with an electrical pressure transducer and a digital acquisition system. For
a given thermodynamic regime (e.g., adiabatic), the value of B or B/V0 should not
change as a function of leak rate, so any convenient leak rate can be used in
performing the calibration. B can, however, vary with temperature, so these
measurements should not be made until the temperature changes in the pipeline
are less than 0.01°C. In general, an 8- to 12-h waiting period will ensure that the
46
-------
temperature changes are small. The selected leak rate should be as large as
possible while still allowing pressure measurements to be made to within 1 psi
and volume measurements to be made to within 1 ml. In most pipelines the total
volume of product that will be drained as the pressure drops from 20 psi to near 0
psi ranges from 20 to 200 ml.
The pressure-volume measurements can be difficult to make from an operational
standpoint if the leak rate is too large. In general, it takes two people if pressure
measurements are made with a mechanical gauge and the cumulative volume of
released product is read in a graduated cylinder. The best way to make this
measurement is to read the pressure in predetermined intervals of 5 or 10 ml as
the graduated cylinder is filling up with product that is draining from the line. For
most pipelines, accurate measurements can be made if the leak-making
apparatus is set to allow a flow rate of between 0.20 and 0.5 gal/h at the test
pressure; the exact flow rate of the leak is unimportant and does not need to be
measured. The data collection should be completed in less than 2 min; if the test
is completed in less than 2 min, the value of B should be nearly equal to the
value of B for an adiabatic process. Enough pairs of pressure-volume data
points should be collected so that the slope of the line can be accurately
determined. It is recommended that at least five points be used. Three
measurements of B/V0 should be made and the median value should be
reported. The differences between the median value and the minimum and
maximum values should be less than 10%.
The volume of the product in the pipeline can be estimated if the diameter and
length of the pipe and fittings are known. An estimate can be made from final
construction drawings that show what was actually installed. The volume of the
product in the pipeline should be known to within 1 gal (the amount of product
contained in a 6-ft length of 2-in.-diameter pipe, or 10% of the total volume in the
line).
4.4 PRODUCT TEMPERATURE
It is very difficult to measure the rate of change of temperature of the product inside a
pipeline. To do this would require an array of temperature sensors capable of
measuring the rate of change of temperature to 0.2°F. Since two to three uniformly
spaced sensors are required for each 10 gal of product in the line, a 100-ft, 2-in.-
diameter line would require approximately six temperature sensors. Even if such an
array measured the product temperature accurately, there would be no guarantee of
standardized evaluation conditions. This is because the temperature of the product
in the pipeline changes exponentially over time and the rate of change depends on
the heat transfer properties of the pipeline and the backfill and soil surrounding it, as
well as on the temperature of the product in the pipeline and the temperature
47
-------
distribution in the backfill and soil at the start of the test. When the dispenser is
turned on, product from the bottom of the storage tank, which is at a certain
temperature, is pumped through the pipeline, whose surrounding backfill and soil
may be at a different temperature. As more product is dispensed through the
pipeline, the temperature distribution in the surrounding backfill and soil changes.
Thus, the temperature of the backfill and soil immediately surrounding the pipeline
may be very different from the temperature of the soil some distance away. The
degree of difference depends on how often product was dispensed prior to the test
and how long it has been since the last dispensing of product through the pipeline.
As a consequence, the actual rate of change of temperature of product in the
pipeline during two leak detection tests can be very different, even though the
temperature difference between the product in the tank and the temperature of the
backfill or soil located far away from the pipeline is the same. Heat-transfer
calculations with mathematical models and experimental measurements on LIST
pipeline systems suggest that the rate of change of product temperature will
decrease to less than 0.02°F/h (0.01 °C/h) 8 to 12 h or less after dispensing has
ceased. Therefore, a leak detection system whose protocol includes a waiting period
between the last dispensing of product and the beginning of a test will always
experience more benign temperature conditions than a system whose protocol does
not require a waiting period. Simply comparing the temperature difference between
product at the bottom of the tank and product in the pipeline (or the ground
temperature at the same depth as the pipeline but not adjacent to it) is not sufficient,
because this difference does not accurately account for the distribution of
temperature in the backfill and soil.
Figure 4.5 illustrates the difference in the rate of change of temperature of the
product within the pipeline under two different ground conditions. In Figure 4.5(a),
the temperature of the ground is constant and in Figure 4.5(b), the temperature of
the ground changes as the distance from the pipe increases. In Figure 4.5(b) the
initial ground temperature was the same as in Figure 4.5(a); product that was 9°F
(5°C) warmer than the ground was then dispensed continuously through the pipeline
for 16 h. Figure 4.5(c) shows the rate of change of temperature under both ground
conditions; in this instance the temperature of the product at the bottom of the tank
was 9°F warmer than that of the ground 12 in. from the pipeline. The rate of change
of temperature is clearly different. When there is no dispensing of product through
the line, the initial rate of change of temperature is great, but the temperature of the
product in the pipeline approaches the temperature of the ground more quickly.
This, however, is not typical of what occurs at a retail station. Calculations with a
mathematical model show that the rate of change of temperature (of the product) is
similar regardless of whether product has been dispensed through the line for 1 h or
for 16 h. However, when product has been flowing through the line for only several
minutes, the rate of change is quite different.
48
-------
4.4.1 Equipment and Instrumentation for Generating Product Temperature
In answer to the problem of characterizing temperature conditions, a procedure
has been developed that can be used to ensure that all evaluations of pipeline
leak detection systems are conducted under similar conditions. Four
temperature sensors having a precision and a relative accuracy of 0.2°F are
required. The relative accuracy can be determined by calibrating all four
temperature-sensors together in the same temperature bath so that each is
referenced to the same temperature; in this way differences in sensor readings
can be accurately measured and accounted for.
49
-------
a
U4
D
<
CC
g
^_
LU
o
w
UJ
yj
IT
O
UJ
D
4
UJ
CC
C
UJ
£L
2
yj
19,5
:7,5
15 51=
19.5f
17.5
15,5
0.00 0.05 0.10 0,15 0,20 0,25
- M
21.5
19,5
17,5
s
is.sL
0
12 24
TiME-H
38
Figure 4.5. Product temperature changes predicted for different dispensing operations
using a beat transfer model: (a) temperature of the backfill and soil is constant, (b)
temperature of the backfill and soil that is produced by circulating product through the
pipeline for 16 h at a temperature that was initially constant and 9°F higher than the
backfill and soil, (c) time history of the product temperature changes in the pipeline for the
initial ground conditions shown in (a) and (b).
50
-------
As shown in Figure 4.6, three sensors should be positioned in the ground
somewhere near the midpoint of a 2-in.-diameter pipeline and located 2, 4, and
12 in. away from the outside edge of the pipeline. The most distant temperature
sensor is intended to measure the ground temperature at a location that is not
significantly influenced by the product in the pipeline. If the temperature sensors
are too close to the dispensing end of the pipeline, their readings could be
adversely influenced by ambient air temperature or convective mixing from
product in the vertical extension of the pipe leading into the dispenser. It is
therefore recommended that the sensor array be located at least 5 ft into the line
from either the dispenser or the tank. This may not be possible at an operational
LIST facility that is being used as an instrumented test facility. If there are
multiple pipes in the backfill, it is preferable to use only the outer pipe. The fourth
sensor should be located in the tank, approximately 4 in. from the bottom (or in
whatever container is used to store the product pumped into the pipeline during a
test); this provides an estimate of the temperature of the product that is pumped
from the tank into the pipeline.
T
2 in.
4 in.
12 in.
Figure 4.6. Geometry of the temperature measurements to be made in the backfill and
soil surrounding an underground pipeline.
The temperature sensors should be calibrated before each evaluation, or more
frequently, if required. Calibration is done by inserting the temperature sensors
in a water bath that is continuously being mixed and simultaneously recording the
output of these sensors and a reference sensor. The precision of the reference
51
-------
sensor should be 0.02°F. The accuracy of the reference sensor need only be
good to the nearest 1°F. Calibration data should be obtained in increments of 5
to 10°F or less over the range of ground and product temperatures to be
encountered during the evaluation; a calibration starting at 35°F and ending at
90°F would suffice. At least five points are required to complete the calibration.
A calibration curve is generated by fitting a regression line to the temperature
measured by each sensor being calibrated (y axis) and the temperature of the
water bath from the reference sensor (x axis). The precision of each temperature
sensor is estimated from the standard deviation of the ordinate (y axis). The
accuracy of each temperature sensor is estimated from the intercept of the curve.
It is not essential that the absolute accuracy of each sensor be known, but rather
that each temperature sensor measure the same value. The relative accuracy is
determined from the standard deviation of the intercepts of each calibration curve
or from the standard deviation of a given temperature calculated from each
calibration curve.
4.4.2 Measurement of Product and Ground Temperatures
The temperature conditions in the pipeline during a test must be characterized.
The procedure used to characterize the temperature conditions varies slightly
depending on the type of facility being used: a specialized test facility, one or
more instrumented operational LIST facilities, or several operational LIST facilities
that are not instrumented. When temperature conditions are generated at an
instrumented test facility, product is taken from the bottom of the tank, pumped
into the line, and circulated continuously there for one hour. This serves three
purposes: (1) to produce a difference in temperature between the product in the
pipeline and the surrounding backfill and soil, (2) to produce a temperature
distribution in the surrounding backfill and soil that is similar to that produced by
dispensing product at operational LIST facilities, and (3) to produce repetitive
temperature conditions from test to test. The end of the hour marks the start of a
leak detection test or an initial waiting period. At an instrumented operational
LIST facility, a leak detection test should be initiated at the end of the day
immediately after dispensing operations have ceased. The one-hour circulation
period is then not required, since dispensing of product during normal business
hours has the same effect on the temperature of the backfill and soil (and
therefore on the rate of change of product temperature) as circulating the product
does. Before a test is begun, however, the entire contents of the line must be
flushed for 5 min with product from the bottom of the tank to produce the
temperature condition. When five or more noninstrumented operational LIST
facilities are used, product is, as with the instrumented operational facility,
already adequately mixed, and the test may begin after dispensing operations
have ceased or at the close of the business day.
52
-------
Model calculations suggest that the rate of change of temperature of the product
in the pipeline depends on the distribution of the temperature of the backfill and
soil immediately around the pipeline even though the difference in temperature
between (a) the product in the pipeline and (b) the soil thermally undisturbed by
the pipeline is the same. One could produce a temperature condition by
circulating product through the pipeline for 5 min, and then start a test; however,
to ensure repetitive conditions, one would have to wait 8 h after the test before
producing another temperature condition.
The temperature condition for a particular test is calculated from the following
formula
(4.3)
where
AT = difference between the temperature of the product at the
bottom of the tank and a weighted average of the
temperature of the ground surrounding the pipeline
TTB = temperature of the product 4 in. from the bottom of the tank
or the temperature of the product to be circulated through
the pipeline
TG = [((Ti/3) + (2T2/3))/3] + [2T3/3] = weighted average of the
temperature of the ground surrounding the pipeline
TI, T2, T3 = temperature of the backfill or soil measured 2, 4, and 12 in.
from the outer wall of the pipeline
This equation accounts for the insulating effect of the ground around the pipeline
and the effect of the temperature of the undisturbed ground.
4.5 TRAPPED VAPOR
The pipeline used in the evaluation should be free of any trapped vapor. The
sensitivity of the leak detection system to vapor can be assessed by trapping a
known volume of vapor in the pipeline and conducting one or more leak detection
tests. A simple device has been developed to do this.
4.5.1 Equipment and Instrumentation for Generating Trapped Vapor
Vapor can be trapped in a pipeline system by means of the vapor pocket
apparatus shown in Figure 4.7. This apparatus can be constructed from
common materials that can be purchased at any hardware store. The apparatus
53
-------
consists of a 1.5-in.-diameter tube that has a volume of approximately 100 ml;
the device used in our experiments had a volume of 6.4 in.3 (105 ml). The tube is
capped at the top and bottom and has two valves that can be opened and closed
manually. The volume of vapor trapped in the line nominally depends on the
length of the tube. Table 4.2 gives the volume of trapped vapor in the device as
a function of pressure. The diameter of the tube can be other than 1.5 in.
providing that the volume of the container at zero pressure is greater than 100
ml.
To measure the volume of the container we submerge the vapor pocket
apparatus in water and then close both valves. After removing any excess water
from the inlet or outlet tubes, we can measure the volume of the water in the
container by emptying it into a graduated cylinder and taking a reading of the
level to the nearest 1 ml.
1O
—MANUAL OUTLET VALVE
MANUAL tfMLET VALVE
„--*
-CONNECTOR
Figure 4.7. Mechanical device for trapping vapor in a pipeline system.
54
-------
Table 4.2. Volume of Trapped Vapor in a Tube 1.5 in. in Diameter and 3.5 in. in Length
as a Function of Pipeline Pressure
Line Pressure
(psi)
0
5
10
15
20
25
30
Container Volume*
(ml)
105.2
78.5
62.6
52.1
44.6
38.9
34.6
(in.3)
6.42
4.79
3.82
3.18
2.72
2.38
2.11
* Assuming that atmospheric pressure is 14.7 psi.
To measure the volume of the container we submerge the vapor pocket
apparatus in water and then close both valves. After removing any excess water
from the inlet or outlet tubes, we can measure the volume of the water in the
container by emptying it into a graduated cylinder and taking a reading of the
level to the nearest 1ml.
The entire vapor pocket apparatus must be air-tight. We can check this by
spraying a solution of soapy water at all joints- when the device is under pressure
and looking for bubbles.
The vapor pocket apparatus can be attached to any part of the pipeline while
both the inlet and outlet valves are closed. Once the apparatus is attached to the
line, the outlet valve should be opened to release any residual air that may have
been trapped. Then the outlet valve is closed and the inlet valve is opened to
allow product from the pipeline to enter the container and pressurize it. When
the inlet valve is open, a known volume of vapor is trapped in the line. The
volume of trapped vapor will depend on line pressure. The vapor pocket
apparatus should be insulated during the measurements.
4.5.2 Measurement of Trapped Vapor
The presence of trapped vapor in a pipeline can be identified from the pressure-
volume data collected for estimating the bulk modulus of the pipeline system. As
shown in Figure 3.1, the pressure-volume curve, which can be used to estimate
B/V0 for the pipeline system, is linear in the absence of any vapor in the line.
Curvature suggests the presence of trapped vapor. The volume at zero pressure
is known. If the pressure-volume relationship for vapor is known, the volume of
the trapped vapor in the device can be estimated. It is not necessary to calculate
and report the volume of the trapped vapor if this device is used. The volume of
vapor trapped in the device can be estimated from the following equation of state
for a gas
55
-------
(4.4)
where P! and V^ are the absolute pressure and volume of the vapor in the line at
one pressure, p2 and V2 are the absolute pressure and volume of the vapor in the
line at a second pressure, and n is the gas constant (assumed to be 1.0).
Because of the discontinuity in the pressure-volume curve exhibited in the
absence of any vapor (see Figure 3.1), this relationship cannot be easily used if a
mechanical line leak detector is present in the line.
56
-------
SECTION 5
SELECTION OF EVALUATION CONDITIONS
This protocol requires that the performance of the system be estimated under a wide
range of temperature conditions and leak rates. Section 5 describes how to select the
temperature of the product in the pipeline and the size of the leak in the line. The
conditions selected for an evaluation should reflect the actual conditions under which the
system will be used in the field.
5.1 TEMPERATURE CONDITIONS IN THE PIPELINE
All dispensing through a pipeline should be terminated during a leak detection test on
that line. Dispensing through other pipelines buried in the same backfill and in close
proximity to the pipeline being tested (i.e., within 12 in. of it) should also be
terminated. This is because the temperature of the product in adjacent pipelines can
affect the rate of change of the temperature in the pipeline being tested.
Table 5.1 summarizes the number of tests that must be done for each of the nominal
conditions for which histograms must be generated. A nominal temperature
condition is defined by Eq. (4.3) and requires that product from the tank be
dispensed through the pipeline for 1 h or longer. It is assumed that the temperature
conditions within the range of each 10°F increment will be as uniformly distributed as
possible. This is particularly important for the conditions centered on 0°F; about half
of the conditions should be positive and about half should be negative.
Table 5.1. Number of Tests Required for Each Range of Temperature Conditions
Number Percentage
of Tests of Tests Range of AT (°F)
1 4 AT < -25
4 16 -25 < AT <-15
5 20 -15 +25
*AT is the temperature difference between the ground and the product in the tank estimated
from Eq. (4.3).
At an instrumented test facility, temperature conditions can be created by warming or
cooling the product to be circulated through the pipeline. The following procedure
should be used when multiple temperature conditions are generated during any one
day. As a general rule, the temperature difference between the ground and the
product circulated through the pipeline should change in only one direction. Figure
57
-------
5.1 illustrates a set of three temperature conditions generated over the course of one
day. The backfill and soil are initially at the same temperature as the product added
to the pipeline. Model calculations indicate that these same conditions can be
regenerated on successive days providing that the ground is not subject to different
ambient temperature effects. In the example in Figure 5.1, the test took 2 h. When
accompanied by a 1-h circulation period, the minimum amount of time necessary to
complete each test is 3 h. Three tests can be completed in a 9-h day. All of the
tests for which temperature differences are positive should be done first. A period of
12 h or longer, during which no product is dispensed through the pipeline, should be
allowed to elapse before the negative-temperature tests are begun. It is acceptable
to increase the temperature of the product that is circulated through the pipeline in
equal increments with respect to the initial temperature difference between this
product and the ground 12 in. from the pipeline (i.e., T3); however, the reported AT is
calculated by means of Eq. (4.3) from the temperature measurements made before
the circulation is started. Table 5.2 presents a testing protocol in which three
temperature conditions are produced each day; for this example it is assumed that T3
= 60°F. This test sequence is a good example of one that satisfies the general test
matrix given in Table 5.1. Three vapor pocket tests, which satisfy the criteria
presented in Section 3, are included at the end of Table 5.2. These tests, denoted
by an asterisk, are included at the end of Table 5.1 to better illustrate this example of
a temperature matrix. In an evaluation, the three trapped vapor tests should be
randomly distributed in the test matrix. Assuming three to six tests per day, the
temperature conditions can be generated by circulating product at the temperature
given by TTB and calculating AT from Eq. (4.3). The temperature conditions that
result should satisfy the test matrix in Table 5.1.
The temperature conditions can be also be randomly generated, but it is important
that the absolute value of the tank/ground temperature differences on any given day
of testing always increase or always decrease. For example, any of the following
three sets is acceptable: (1) -2.5, -5.0, -7.5°F, (2) +10.0, +12.5, +15.0°F, or (3) -2.5,
-15.0, -20.0°F. On the following day a set of temperature conditions with a different
sign can be used providing that at least 12 h have elapsed since the last test. It is
not acceptable to both increase and decrease the temperature condition during the
course of a single day (e.g., -2.5, +2.5, -5.0°F). A detailed procedure for randomly
selecting temperature conditions so that they satisfy the above criteria is complex
and unnecessary. If this is to be a blind test, the temperatures can be placed in any
order providing that the above daily criteria are met.
If an instrumented operational LIST facility (Option 2) is being used to evaluate the
leak detection system, it is unlikely that more than one temperature condition can be
generated on any one day. The temperature condition will depend on the product in
the pipeline and the temperature of the ground. Unless there is a way to change the
58
-------
temperature of the product brought from the tank in to the pipeline, an evaluation
performed at an operational facility will take significantly longer than one performed
at a special test facility (Option 1). Enough tests must be conducted to satisfy the
test matrix given in Table 5.1. The time required to collect these data can be
reduced if more than one operational facility is used, particularly if the facilities have
sufficient geographical separation to have different climates during a given season
(e.g., Miami, Florida, and Chicago, Illinois, during the winter). With more than one
instrumented operational LIST facility, a larger range of temperature conditions will
be encountered over a shorter period of time. The total time required to complete
this type of evaluation may be one to six months. Measurement of temperature is
not required if Option 3 or 4 is chosen.
59
-------
21.5
LLJ
Q
<
IT
3
h-
Z
UJ
O
CO
UJ
UJ
cr
Q
LLJ
Q
UJ
cr
ZJ
h-
<
QC
UJ
CL
2
UJ
19.5
a
b
12
TIME-H
24
Figure 5.1. Model predictions of the temperature changes that occur in a pipeline (a) after a
1-h circulation period and (b) after a 5-min circulation period for an initial temperature
difference between the product circulated through the pipeline and the backfill and soil of
4.5°F. The temperature of the product for each circulation period (i.e., each test) was
increased by 4.5°F.
60
-------
More than 25 tests may be needed in order to complete the test matrix of
temperature conditions shown in Table 5.1. Unlike a special test facility, where
temperature conditions can be controlled, operational facilities require more time for
evaluators to gather the required number of tests. It is difficult to acquire exactly the
specified number of tests in each temperature category. Inevitably some categories
will contain a larger percentage of tests than others. For example, 20 tests rather
than 5 may be conducted under the most benign condition, yet only one at each of
the most extreme conditions, changing the relative percentage of tests in each of the
seven categories in Table 5.1. This presents a problem because it produces a
biased performance estimate. Assuming that performance declines as the
temperature condition becomes more extreme, it is likely that the estimate of
performance obtained from this 40-test sequence would be better than it would have
been if the 25-test sequence in Table 5.1 had been followed exactly. Even if the total
number of tests exceeds 25, there are still ways to avoid biasing the results. One is
to randomly select test results in each category until the required number or
percentage of tests is obtained. Similarly, the matrix can be based on the category
with the largest percentage of tests if all the other categories are proportionally
increased by means of a random selection within the category. The relative
percentage of tests should be as it appears in Table 5.1. This means that in some
categories the same test results may be used more than once. Either approach,
then, avoids bias in test results in situations when more than 25 tests are needed to
complete the test matrix given in Table 5.1. The latter approach has the advantage
of using all the data. One might be tempted to try to avoid bias by using the first n
results in each category, discarding any results obtained from tests beyond the
required number; unfortunately this approach itself could bias the results if all the
data from each category were obtained from one end of the category range.
There are many methods that can be used to randomly select the required number of
test results from each category. One method is to use a random drawing procedure.
One way of doing this is as follows: (1) assign a number to each test result in the
category (1 through n); (2) write each number on a piece of folded paper and place
these in a bowl; (3) blindly select pieces of paper from the bowl until the required
number (or percentage) of tests for that category is obtained. A computerized
version of this procedure could also be used.
61
-------
Table 5.2. Recommended Procedure for Generating a Temperature Condition at an
Instrumented Test Facility
Test Number TTB T^ AT
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26*
27*
28*
60
62
64
66
68
70
72
74
77
79
81
83
86
Wait 12h or longer before
58
56
54
52
50
48
46
43
41
39
37
34
74
74
74
60
60
60
60
60
60
60
60
60
60
60
60
60
proceeding with test matrix
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
0
2
4
6
8
10
12
14
17
19
21
23
26
-2
-4
-6
-8
-10
-12
-14
-17
-19
-21
-23
-26
14
14
14
5.2 INDUCED LEAK RATES
It is desirable to perform, if possible, more than one leak detection test under each
temperature condition, because this will reduce the amount of time necessary to
complete an evaluation. If circumstances permit the generation of more than one
temperature condition in a single day, the noise histogram can be generated from a
test on a nonleaking line and the signal-plus-noise histogram can be generated from
a test on a line leaking at the EPA-specified rate and possibly at other leak rates too.
This is sometimes difficult to do, because the temperature of the product can change
significantly from one measurement period to another, even though these
measurements are closely spaced. Generally, the guidelines for closely spaced
multiple leak detection tests at different leak rates under a given temperature
condition, including a test on a tight line (i.e., a leak rate of 0.0 gal/h), are as follows.
62
-------
• Up to three tests can be conducted if the leak detection system's test
protocol requires a waiting period between the last input of product into the
pipeline and the start of the data collection period, if this waiting period is
greater than 6 h, and if the duration of the data collection period for each test
is 1 h or less.
• Up to three leak detection tests can be conducted if the duration of the data
collection period is less than 20 min, regardless of the length of the waiting
period.
• Up to two leak detection tests can be conducted if the waiting period is
greater than 4 h and the duration of the data collection period for each test is
1 h or less.
• Up to two leak detection tests can be conducted if the duration of the data
collection period is less than 30 min, regardless of the duration of the length
of the waiting period.
If multiple tests are conducted under the same temperature condition, the order of
the leak rates should be randomly selected. This is important because the rate of
change of temperature decreases with time and the test results would be biased if
data for the same leak rate were always collected first. If one of these criteria for
multiple tests cannot be satisfied, a new temperature condition must be created for
each leak rate.
There are two types of testing scenarios: the test crew can have full knowledge of
the conditions, or they can be placed in a blind testing situation.
5.2.1 Known Test Conditions
In the first scenario, the temperature and leak conditions are known by both the
testing organization (the manufacturer of the leak detection system) and the
evaluating organization. This scenario includes tests at a minimum of two leak
rates, 0.0 gal/h and the EPA-specified leak (i.e., 0.1, 0.2, or 3.0 gal/h) for which
the performance is to be determined. If the relationship between the signal-plus-
noise provided by the manufacturer can be verified experimentally, it can be used
to generate a signal-plus-noise histogram. (The way to do this is to shift the
noise histogram appropriately.) It is then not necessary to conduct tests at the
EPA-specified leak rate. The temperature conditions will be generated from
smallest to largest, first for positive tank/ground temperature differences and then
for negative. If the requirements for multiple testing are satisfied, the order of the
leak rates for each test under a given temperature condition should alternate;
there must be a total of 25 temperature conditions, satisfying the general range
of conditions in Table 5.1. If the requirements for multiple tests are not satisfied,
63
-------
only one test can be done under each temperature condition, and the number of
temperature conditions that must be generated doubles to a total of 50. Table
5.3 presents a set of suggested test conditions for two leak detection tests per
temperature condition. Table 5.3 is based upon the temperature conditions in
Table 5.2. Leak rates of 0.0 gal/h (required to generate the noise histogram) and
0.1 gal/h (required to generate the signal-plus-noise histogram for the line
tightness test specified by the EPA regulation) are used. The three trapped
vapor tests are also included at the end of the test matrix and are denoted by an
asterisk. In an evaluation, these tests should be randomly distributed throughout
the test matrix.
64
-------
Table 5.3. Example of Test Conditions When More Than One Test Can Be Done for a
Temperature Condition
Test
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26*
27*
28*
TTB
(°F)
60
62
64
66
68
70
72
74
77
79
81
83
86
Wait 12
58
56
54
52
50
48
46
43
41
39
37
34
74
74
74
T3
(°F)
60
60
60
60
60
60
60
60
60
60
60
60
60
h or longer before
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
AT
(°F)
0
2
4
6
8
10
12
14
17
19
21
23
26
proceeding
-2
-4
-6
-8
-10
-12
-14
-17
-19
-21
-23
-26
14
14
14
Leak No. 1
(gal/h)
0.1
0
0.1
0.1
0.1
0
0.1
0
0.1
0
0
0
0.1
with test matrix
0.1
0
0.1
0.1
0
0
0
0.1
0.1
0.1
0.1
0
0
0
0
Leak No. 2
(gal/h)
0
0.1
0
0
0
0.1
0
0.1
0
0.1
0.1
0.1
0
0
0.1
0
0
0.1
0.1
0.1
0
0
0
0
0.1
0.1
0.1
0.1
5.2.2 Procedures for Blind Testing
Full-scale blind testing is not possible because the test crew knows, from this
protocol, what leak rates are used in the evaluation and that the temperature
conditions will be systematically increased or decreased. However, they do not
know the order in which the leaks will be generated, and they do not know what
temperature condition is being used; partially blind tests are therefore possible.
There are two types, described below as Procedure 1 and Procedure 2. In
65
-------
Procedure 1 there are two leak rates, one of which will be zero and the other of
which will be the EPA-specified leak rate. The same number of tests (usually 25)
is conducted at each leak rate. Then there are three to five additional tests under
unknown temperature conditions and at an unknown leak rate. In Procedure 2,
there are four leak rates: 0, 0.05, 0.1, and 0.2 gal/h for monthly monitoring and
line tightness tests and 0, 2.5, 3.0 and 3.5 gal/h for hourly tests. Twenty-five
tests are conducted at the zero leak rate and 25 at either the 0.1 or 0.2 rate (the
EPA-specified leak rates). Then an additional 13 tests are conducted at 0.1 or
0.2, and an additional 12 at 0.05. This brings the total number of tests to 75. In
either procedure, as long as the test crew must report the results of one test
before going on the next, blind testing will be assured. Any other procedure, or
variation on the two procedures below, is acceptable if the same conditions for
blind testing are met.
5.2.2.1 Procedure 1
Like the known-condition scenario, Procedure 1 requires that for each
temperature condition at least 25 tests be conducted at a leak rate of 0.0
gal/h and at the EPA-specified leak rate. The difference between Procedure
1 and the known-condition scenario is that in the former the order of the leak
rates per temperature condition is randomly selected, and between three and
five additional temperature and leak conditions are introduced. These three
to five additional tests, which will not be included in the performance analysis,
will represent temperature differences greater than ±15°F between the tank
and the ground, split as evenly as possible between positive and negative,
and leak rates between 0.1 and 0.5 gal/h. In these extra tests, the nature of
the temperature conditions, the size of the leaks, and even the number of
tests are unknown to the test crew. The additional three to five tests tend to
give large temperature- and leak-induced flow rates that might make the
results look anomalous, and the test crew may be tempted to reject the
results of such tests. However, if for any reason other than an obvious
malfunction of the test equipment identified during the test itself the test crew
declares one of these tests invalid, the evaluation should not be considered a
blind test.
A random number generator or a random drawing of conditions can be used
to select the number of extra tests, the leak rates, and the temperature
conditions for each test. Or, the number of extra tests can be determined by
writing the numbers 3, 4, and 5 on pieces of folded paper, placing these in a
bowl, and randomly drawing one of the numbers. This can also be done by
66
-------
multiplying the output of a random number generator* (for example, a
computer spreadsheet, scientific calculator or statistical tables) by five,
rounding to the nearest integer value, and using the first number that is 3 or
greater. Once the number of tests has been determined, a temperature
condition must be selected for each test. If the number is odd, the extra
temperature condition should be positive. Select one to three temperature
conditions from the positive tank/ground temperature differences in the range
+15 to +30°F, and randomly select one to three temperature conditions from
the negative tank/ground temperature differences in the range -15 to -30°F.
This can be done by randomly drawing temperature conditions, differing by
increments of 1°F, from a container. This can also be done by multiplying the
output of the random generator by 15, rounding to the nearest integer, and
adding +15°F to get the positive temperature differences and subtracting -
SOT to get the negative. Each leak rate in gallons per hour can be
determined with a random number generator by dividing the output of the
random number generator by 2.5 and adding the result to 0.1. Alternatively,
twenty random selections of leak rates between 0.1 and 0.5 gal/h are given in
Appendix D. Randomly select a number between 1 and 20 from a container
to determine which table to use, and then select as many leak rates as there
are tests to be conducted.
Acceptable methods of randomly selecting the leaks for the 25 tests under
each temperature condition are as follows. If the criteria for conducting two
tests under each temperature condition are met, the first and second leak
rates for each temperature condition can be determined in the following way.
Place two pieces of folded paper in a bowl, each piece of paper having one of
the leak rates written on it, and randomly draw a number for each
temperature condition; at least one of the leak rates for each temperature
condition should be the EPA-specified leak rate. If a random number
generator is used, its output for each leak can be rounded off to the nearest
integer, with 0 being a leak rate of 0.0 gal/h and 1 being a leak rate equal to
the EPA-specified leak rate. If the multiple-test conditions are not satisfied, a
random drawing should be made for each temperature condition.
5.2.2.2 Procedure 2
Procedure 2 requires that for each of the 25 temperature conditions, the
following tests be conducted: 25 leak detection tests at a leak rate of 0.0
gal/h; 25 leak detection tests at the EPA-specified leak rate, and 12 and 13
tests, respectively, at two other leak rates. Table 5.4 suggests the leak rates
It is assumed that any output of the random number generator is between 0.0 and 1.0.
67
-------
recommended for making an estimate of performance for the system at the
EPA-specified leak rate.
Table 5.4. Recommended Leak Rates for Procedure 2
Number of Tests Conditions for EPA-specified Leak Rates - gal/h
0.1 0.2 3.0
25
25
13
12
0.0
0.1
0.2
0.05
0.0
0.2
0.1
0.05
0.0
3.0
2.5
3.5
The first step in generating a matrix of leak conditions is to select the EPA-
specified leak at which the leak detection system is to be evaluated. Each
test run will use three out of four possible leak rates. These leak rates should
be randomly selected for each temperature condition. A table of conditions
can be generated as illustrated in Table 5.5. The leak rates for each test can
be randomly selected by writing each of the four leak rates on a piece of
paper, placing the folded pieces of paper in a container, and randomly
drawing three leaks for each test. If a random number generator is used,
assign a number between 1 and 4 to each of the four leak rates, multiply the
output of the random number generator by 3, round to the nearest integer
and add one to the integer. Three different numbers should be generated for
each test.
Table 5.5. Illustration of a Possible Test Matrix for Evaluation of a Leak Detection
System at 0.1 gal/h
Test Temperature Condition Leak Rate 1 Leak Rate 2 Leak Rate 3
No. (°F) (gal/h) (gal/h) (gal/h)
1 2.5 0.1 0.0 0.05
2 5.0 0.0 0.2 0.1
3 7.5 0.2 0.0 0.1
23 -20.0 0.0 0.1 0.2
24 -22.5 0.0 0.2 0.05
25 -25 0.2 0.1 0.0
Procedure 2 can be used to generate a performance estimate of the system
at leak rates other than the EPA-specified leak rate. To do this, use the test
results from the two alternative leak rates to generate a signal-plus-noise
histogram. If the output of the leak detection system is reported, use the
noise histogram, shifted by the difference between the mean of the noise
histogram and the mean of each of the two alternative leak rates, to generate
68
-------
a signal-plus-noise histogram for each of these two leak rates. The signal-
plus-noise histogram can be used directly for each of these two alternative
leak rates if at least 25 tests have been conducted at each leak rate. To
satisfy this test matrix, a total of 38 tests would have to be conducted to
obtain 25 test results for each of the four leak rates. If the leak detection
system does not output and report a measured quantity, but instead uses a
preset threshold, a total of 25 tests is required for each leak rate at which a
performance estimate is desired.
69
-------
SECTION 6
EVALUATION PROCEDURE FOR SYSTEMS
THAT REPORT A FLOW RATE
Some leak detection systems measure an output quantity and compare it to a
predetermined threshold to assess whether the pipeline is leaking. If the measured
quantity is less than the threshold, the pipeline is declared tight. Otherwise, the pipeline
is either declared leaking or another test is conducted to confirm or refute the results of
the first. Other systems use a preset-threshold switch that is activated only if the
changes in the line are large enough; no quantity is reported. The protocol for
evaluating systems that measure and report the output quantity is described here in
Section 6. The protocol for evaluating systems that use a preset-threshold switch is
presented in Section 7. The procedure for evaluating both types of pipeline leak
detection systems consists of the same general sequence of steps presented in Section
3.3. There are, however, slight differences in estimating the performance characteristics
of the two types of systems and in how to analyze the noise and signal-plus-noise data
to derive a performance estimate in terms of PD and PFA-
6.1 PERFORMANCE CHARACTERISTICS OF THE INSTRUMENTATION
Before performing any evaluation experiments with a leak detection system, it is
necessary to ensure that the system is working correctly and is properly calibrated.
An uncalibrated system could produce unexpected and sometimes meaningless
results. In addition to this data quality assurance, the calibration also provides a
measurement of the precision and accuracy (or bias) of the system's sensor(s).
While these measurements may not necessarily be used quantitatively in the
calculation of the PD and PFA of the system, they are used qualitatively to determine
the advisability of proceeding with the evaluation. If the instrumentation (or system)
noise is so large that the required performance could not be achieved even if no
other noise sources were present, the evaluation procedure could be stopped, and a
reassessment of the system design might be considered.
Each sensor used by the leak detection system should be calibrated in a controlled
environment to determine what is being measured (i.e., specificity) and to make an
estimate of the resolution, precision, accuracy, minimum detectable signal, and
response time. For most instruments that measure a physical quantity (for example,
volume, pressure, or temperature), the specificity is obvious. The resolution of the
system is the smallest division for which a quantity is measured; since the resolution
is usually well known, it does not have to be measured as part of this protocol, but it
does have to be reported. The minimum detectable quantity is defined in this
protocol as that quantity that can be detected with a PDof 0.95 and a PFA of 0.05;
70
-------
assuming that the instrumentation noise is normally distributed, the minimum
detectable signal is 3.3 times larger than the precision.
An estimate of the precision and accuracy of each instrument should be made
against a reference standard. This can be done by making three measurements of
each of at least six different values of the measured quantity. These values should
encompass the dynamic range of the system or the range of conditions under which
it will be operated. The precision and accuracy are estimated from a regression line
fit to the measured quantity plotted on the y axis and the reference standard plotted
on the x axis. The precision is the standard deviation of the ordinate and the
accuracy is equal to the intercept of the line. (See Appendix E for a description of
how to calculate the mean and standard deviation of a set of measurements and how
to fit a regression line to a set of measurements.)
If a pressure transducer is used to monitor the pressure changes in the pipeline over
a range of 0 to 40 psi, the calibration might be done at nominal intervals of 5 psi
between 0 and 40 psi. Thus, three measurements would be taken at nine known
pressures (e.g., 0, 5, 10, 15, 20, 25, 30, 35, and 40 psi). Performing the calibration
exactly at 5-psi intervals is not essential. The calibration could be done at 1, 5.5, 9.7,
15, 21, 27, 31.5, 35.2, and 40.8 psi. It would also be acceptable to take data at six
pressures in nominal intervals of 8 psi (e.g., 4, 12, 20, 28, 36, and 44).
An estimate of the threshold flow rate, defined at 20 psi, beyond which a leak will be
declared is also required. For leak rates of 0.1 and 0.2 gal/h, the flow rate at which
the threshold will be exceeded should be measured to within 0.015 and 0.030 gal/h,
respectively; for a leak rate of 3.0 gal/h, the flow rate at which the threshold will be
exceeded should be measured to within 0.25 gal/h. This estimate can be made on
the pressurized pipeline system that will be used in the evaluation. (The sources of
ambient noise, for example, the changes in product temperature, should be
minimized while this estimate is being made.) Different leak rates are generated,
from small to large, until the threshold is exceeded.
6.2 DEVELOPMENT OF THE NOISE AND THE SIGNAL-PLUS-NOISE
DATA
In order to calculate the PD and PFA, one must first develop the cumulative frequency
distributions (CFDs) from the histograms of the noise and the signal-plus-noise. As
shown in Figure 2.4, the PD and PFA are derived from these CFDs along with the
detection system's threshold and the leak rate of interest. In cases where the signal
is independent and additive with the noise, the signal-plus-noise CFD is just a replica
of the noise CFD shifted by the amount of the leak rate (as is the case in Figure 2.3).
However, it cannot, in general, be assumed that the signal and the noise are linearly
related. This relationship must be verified experimentally.
71
-------
If the system uses a multiple-test strategy, the histogram of the noise and the
histogram of the signal-plus-noise are generated from that test sequence which was
the basis for declaring a leak. In addition to histograms used to develop a
performance estimate of the system, a second performance estimate is requested.
This second estimate is based only on the results of the first test in the multiple-test
sequence. Refer to Section 3.2.3 for a discussion of multiple-test strategies.
In this protocol, it is assumed that the evaluation is being performed to obtain the PD
and PFA at the leak rate specified in the EPA regulation for the type of system being
evaluated, e.g., 0.1 gal/h for a line tightness test, 0.2 gal/h for a monthly monitoring
test, and 3 gal/h for an hourly test. Thus, the procedure described below leads to the
development of a noise CFD and a signal-plus-noise CFD for the leak rate of
greatest regulatory interest for a line tightness test, a monthly monitoring test, and an
hourly test. If local regulations specify leak rates more stringent than those in the
EPA regulation, the local specification can be substituted for the EPA-specified leak
rate.
Five options for developing the cumulative frequency distribution of the noise and the
signal-plus-noise are described in the following sections. Each option is described in
terms of procedure and data analysis. All require that the histograms be
experimentally determined. The way to do this is to accumulate the results of tests
that cover a wide range of temperature conditions.
6.3 EVALUATION PROCEDURE
The reader will recall, from Section 3.3, the general summary of the steps involved in
the protocol. These steps are reiterated here, in a more specific way, as they apply
to each of the five options. Step 2 of the protocol summarized in Section 3.3
presents the five options for collecting the data necessary to evaluate the
performance of a pipeline leak detection system that measures and reports an output
quantity. Since Step 2 is to choose one of the five options, which has obviously been
done at this point, this step is omitted from procedures described below.
6.3.1 Option 1 - Collect Data at a Special Pipeline Test Facility
In Option 1, data are collected at special pipeline test facility. The histogram of
the noise is generated from the results of actual tests with the leak detection
system on a nonleaking pipeline over a wide range of environmental conditions.
These conditions must include a wide range of product temperature changes.
Option 1 is most easily implemented at a test facility like the EPA 's LIST Test
Apparatus, where the integrity of the pipeline system is known and a range of
environmental conditions can be generated and monitored quantitatively. The
signal-plus-noise histogram for the EPA-specified leak rate can be compiled
either directly from tests with the leak detection system over the same conditions
72
-------
used to generate the noise histogram or from the noise histogram and an
experimentally validated relationship between the signal and the noise.
The test procedure will be applied to a pipeline system that meets the minimum
specifications presented in Section 3.1. Below are the steps that should be
followed to evaluate a leak detection system at a test facility. The steps
correspond to those summarized in Section 3.3. Step 2, which is the selection of
the evaluation option, has been omitted.
Step 1 - Describe the leak detection system to be evaluated. A general
description of the leak detection system must be prepared before the system is
evaluated. Attachment 1 in Appendix B is a form that is provided for this
purpose; additional information can be included if so desired or if such
information is necessary to complete the description of the system. The
description includes the important features of the instrumentation, the test
protocol, data analysis, and detection criterion. The system's test protocol should
be followed during the conduct of the evaluation. If the system uses a multiple-
test strategy to determine whether the pipeline is leaking or not, this same
strategy should be followed during the evaluation.
Step 3 - Select leak rates and temperature conditions. Option 1 requires that
25 leak detection tests be conducted according to the system's testing protocol
on a nonleaking pipeline under temperature-conditions that satisfy the seven
different categories of tank/ground temperature differences given in Table 5.1.
Option 1 also requires that 25 leak detection tests be conducted under the same
range of temperature conditions with a leak equal to the EPA-specified leak rate
if a relationship between the signal and the noise is not known, or if a direct
estimate of performance is desired at this leak rate. A matrix of temperature and
leak conditions must be developed. The matrix depends on how the signal-plus-
noise histogram is to be developed and whether the evaluation is to be done
under conditions that are known or unknown to the test crew. A detailed
description of how to generate a test matrix is presented in Section 5. Option 1
also requires that three tests be done with vapor trapped in the line. (The vapor
pocket device described in Section 4.5 can be used to introduce the vapor into
the line.) Thus, if this option is chosen, the minimum number of leak detection
tests is 28.
Step 4 - Assemble the required equipment and diagnostic instrumentation.
The following equipment and diagnostic instrumentation are required:
leakmaker, pressure sensor, a minimum of four temperature sensors, pipeline
compressibility device, vapor pocket device, graduated cylinders, and stopwatch.
A description of the equipment and how to use it is presented in Section 4.
73
-------
Step 5 - Verify that the line is not leaking. The pipeline system to be used in
the evaluation has to be tight. Before the evaluation is begun, the line should be
tested with a leak detection system that has a known performance. If a test
facility is used, the integrity of the line does not have to be verified before each
evaluation, but this should nevertheless be done at regular intervals. It is
particularly important to verify that the pipeline system is not leaking if a third
party evaluation is being performed. If there is a small leak in the pipeline, the
performance of the system being evaluated will be unnecessarily degraded.
Step 6 - Measure the pipeline compressibility characteristics. The pipeline
used in the evaluation should have a B of 25,000 psi; any system with a B
between 15,000 and 40,000 psi is acceptable. Measurements of B/V0 and B
should be made when the temperature changes are small (i.e., less than 0.02°F
over the duration of the measurement period) and should follow the procedure
given in Section 4.3. The leak detection system should not be physically present
in the line if it affects the magnitude of B or B/V0. Unless temperature sensors
such as thermistors are used to measure temperature in the line, measurements
of B/V0and B cannot be made until the pressure in the line stays within 1 psi over
a period equal to the average duration of a B/V0 measurement (approximately 2
min). Three estimates of B and B/V0will be made and the median value
reported.
If the measured value of B is outside the specified range, the device described in
Section 4.3 can be used to modify the compressibility characteristics of the
pipeline and therefore the bulk modulus. Add the compressibility device to the
pipeline and measure BA/0. Repeat this procedure until B is as close to 25,000
psi as possible or is within the specified range.
The results of these measurements should be tabulated and reported on
Attachment 3 in Appendix B.
Step 7 - Determine the performance characteristics of the instrumentation.
Estimates must be made of (1) the minimum quantity detectable by the system,
(2) the precision and accuracy of each instrument used to collect the data over
the dynamic range required for the measurements, and (3) the response time of
the system. The resolution and flow rate of the threshold in gallons per hour
must be reported. These measurements can be made in a special calibration
unit or on the pipeline system itself when the noise is negligible. The general
procedures required to estimate the performance characteristics of the
instrumentation are described in Section 6.1.
Step 8 - Develop a relationship between the leak and the output of the
measurement system, if necessary. If the relationship between the signal and
74
-------
the noise is known and a direct estimate of the signal-plus-noise histogram at the
EPA-specified leak is not made experimentally, or if the general relationship
between the signal and the noise is desired, the relationship must be verified
experimentally. (This step is not necessary if the test matrix requires 25 tests at
the EPA-specified leak rate.) The two-step procedure for developing this
relationship is described in Section 4.2.3. The test results should be summarized
in the tables in Attachment 7 in Appendix B. The appropriate forms from
Attachments 4 and 5, which describe the temperature and leak conditions, as
well as the test results, should also be completed.
Steps 9 and 10 - Collect the -noise data, the signal-plus-noise data, and the
trapped vapor data. The pipeline leak detector may have been isolated from
the line during the bulk modulus measurements in Step 5. If so, it should now be
reconnected so that the leak detection tests can be conducted. A leak detection
test should be performed according to both manufacturer's protocol and the test
matrix developed in Step 3. The result of each test should be recorded in terms
of the output of the system. The three tests in which trapped vapor is present in
the pipeline are also part of the test matrix and should be included in the overall
data collection effort. There should be break of 12 h or longer between tests
conducted under positive temperature conditions and those conducted under
negative conditions. A temperature condition is created by circulating product
through the pipeline system for 1 h before the test; the temperature of this
product must be different from the temperature of the backfill and the ground
around the pipeline. (The leak rate can be set at any time during this same 1-h
period.) All dispensing through a pipeline should be terminated during a leak
detection test on that line. Dispensing through other pipelines buried in the same
backfill and in close proximity to the pipeline being tested (i.e., within 12 in. of it)
should also be terminated.
The equipment and the procedures for generating a leak in the line are described
in Section 4.2. If possible, all leaks will be generated at the at a line pressure
equal to the pressures specified in Section 4.2 (i.e., 10 psi for hourly testing
systems and 20 psi for all other types of systems). If this cannot be done, the
leak can be generated at another pressure (e.g., the operating pressure of the
line) provided that it is equivalent to leak rates defined in Section 4.2. The leak
rate used in each test should be measured and reported. Once the leak has
been generated, the line pressure can be readjusted, if this is required by the
system's test protocol, to the appropriate pressure for the test.
The result of each test should be recorded in terms of the output of the system.
These results constitute the data needed to build the histograms of the noise and
the signal-plus-noise at the EPA-specified leak rate. If a multiple-testing
procedure is used, noise and signal-plus-noise histograms must be compiled
75
-------
from the data used to determine whether the pipeline is leaking and from the first
test of the multiple-test sequence; refer to Section 3.2.3 for additional details.
The test results should be partitioned into the following groups:
(1) data from the 25 tests at a zero leak rate
(2) data from the 25 tests at the EPA-specified leak rate
(3) data from the tests at any other leak rate
(4) data from the three trapped vapor tests
(5) data from any extra tests
Compute the mean, standard deviation, and 95% confidence intervals on the
means and standard deviations for the data in (1) through (3). The formulas
necessary to perform these calculations are given in Appendix E.
The data in (1) are used to define the noise, and the data in (2) are used to
define the signal-plus-noise at the EPA-specified leak rate. A performance
estimate can be derived directly from cumulative frequency distributions of the
noise and the signal-plus-noise according to the PD/PFA analysis presented in
Section 6.4. Performance estimates can be made at the other leak rates from
the noise data in (1) if the signal-plus-noise data in (3) are sufficient.
A signal-plus-noise cumulative frequency distribution can be generated for any
leak rate if the relationship between the signal and the noise is known and has
been validated experimentally with the data obtained in Step 6. The relationship
between the signal and the noise is used to shift the noise histogram
appropriately.
The temperature and leak conditions and the tests results obtained for these
conditions should be tabulated and reported on Attachments 4 and 5 in Appendix
B.
Step 11 - Sensitivity to Trapped Vapor. The results of the tests on lines with
trapped vapor should be tabulated and reported on the standard form included as
Attachment 6 in Appendix B.
Step 12 - Performance Analysis. The performance of the system can be
calculated from the data partitioned for specific leak rates, PDs and PFAs. The
protocol is designed so that the PD and PFA of the system are established at the
manufacturer's threshold and at the leak rate specified by the EPA regulation
(i.e., 0.1, 0.2, or 3 gal/h) at a test pressure of 20 psi. If the evaluation is not done
at the test pressure specified by the EPA, there is a method with which to
76
-------
calculate an equivalent leak rate at the non-EPA test pressure. So that each
system can be compared to others, Attachment 2 in Appendix B provides tables
for reporting a variety of performance estimates. If the leak detection system
uses a multiple-test procedure, performance estimates should follow the
system's protocol, and histograms should be generated from the data from both
the last test and the first test. The analysis of the performance of a detection
system in terms of PD and PFA is described in Section 6.4.
Step 13 - Evaluation Report. The results of the evaluation are tabulated and
reported in the standard format presented in Appendix A and Appendix B. The
performance characteristics of the instrumentation, the performance estimates of
the system's ability to detect leaks under ambient environmental conditions, and
the sensitivity of the system to trapped vapor will be presented in a standard set
of tables. A leak detection system, as used in the field, meets the EPA standard
for the leak rate specified in the regulation if the calculated PD is 0.95 or greater
and the PFA is 0.05 or less. The temperature and leak rate conditions under
which the system was evaluated should be tabulated and reported along with the
test results for each temperature condition and each leak rate. The report also
includes a general description of the pipeline system that was used in the
evaluation. Finally, a section is provided for general comments.
6.3.2 Option 2 - Collect Data at One or More Instrumented Operational
UST Facilities
In Option 2, data are collected at one or more instrumented, operational UST
facilities. A special test facility (Option 1) has the equipment necessary to
generate different temperature conditions. If this type of equipment is available
at an instrumented operational UST facility, Option 2 is identical to Option 1. If
there is no way to generate different temperature conditions, enough tests must
be conducted to cover the range of temperatures specified in Table 5.1. The
procedure for completing the test matrix so as to avoid biasing the performance
estimate is described in Section 5.1. Other than this, the procedures are the
same for both options.
6.3.3 Option 3 - Coiled Data over a 6- to 12-month Period at 5 or More
Operational UST Facilities
Option 3 is nearly identical to Option 2 except that the tests are conducted on a
limited number of nonleaking, operational UST pipeline systems that represent
the conditions under which the leak detection system will be used. In order to
capture a range of climatic conditions, five different locations are used, each in a
different region of the United States. In order to capture the seasonal effects at
each location, periodic tests of the lines are conducted at intervals of
77
-------
approximately one month over a 6- to 12-month period. In Approach 3, at least
60 tests are needed (12 at each site, conducted at regular intervals). Because
the stations are limited in number, the integrity of the pipeline systems should be
verified, if possible, before the data collection begins. This option is best
implemented when the relationship between the signal and the noise is well
known. In this way, the signal-plus-noise histogram can be characterized without
the need for extensive measurements at one or more of the sites. This option is
particularly suited to automatic systems that routinely conduct a test of the
pipeline whenever the LIST facility doses.
Option 3 comprises the following steps, which correspond to those summarized
in Section 3.3. Many of them are similar to those presented in Option 1. Again,
Step 2 is omitted.
Step 1 - Describe the leak detection system to be evaluated. A
general description of the leak detection system must be prepared before
the system is evaluated. The form included as Attachment 1 in Appendix
B is provided for this purpose; additional information can be included if so
desired or if such information is necessary to complete the description of
the system. The description includes the important features of the
instrumentation, the test protocol, and detection criterion. The system's
test protocol should be followed during the conduct of the evaluation. If
the system uses a multiple-test strategy to determine whether the pipeline
is leaking, this same strategy should be followed during the evaluation.
Step 3 - Select leak rates and temperature conditions. Option 3
requires that a minimum of 12 tests be conducted at each of the five
operational LIST facilities, for a total of at least 60 tests, over a 6- to 12-
month period. The protocol requires that tests be conducted at intervals
of approximately 2 to 4 weeks. They can be conducted more frequently if
the evaluator so desires. A test must meet the following conditions:
• It must be started within 30 min of the last dispensing of product
through the pipeline.
• It must be started within 12 h (preferably within 6 h) of a delivery of
product to the tank.
Only stations that receive, on the average, a delivery of product to the
storage tanks on a weekly basis can be used. Each test should be
conducted as soon as possible after a delivery of product to the tank; this
ensures that the temperature conditions will be approximately the same
as those generated for Options 1 and 2. It is desirable to perform a leak
78
-------
detection test at each available opportunity (even as often as once per
delivery). The date and time of the start and end of each test, the time
that dispensing operations were terminated prior to the test, and the date
and time of the last delivery of product to the tank should be recorded and
tabulated. The nominal operating pressure of each pipeline system used
in the evaluation should be measured and recorded. These data will be
used to generate and interpret the noise histogram. Option 3 does not
require that a set of tests be done at the EPA-specified leak rate and
does not require that trapped vapor tests be conducted. It does require
that the tests used in the performance analysis be conducted under the
temperature conditions specified in Table 5.1. The geographical diversity
of the stations and seasonal effects at each station will serve to satisfy
those temperature conditions.
Step 4 - Assemble the required equipment and diagnostic
instrumentation. The following equipment and diagnostic
instrumentation are required: leakmaker, pressure sensor, graduated
cylinders, and stopwatch. A description of the equipment and how to use
it is presented in Section 4.
Step 5 - Verify that the line is not leaking. The pipeline used at each
operational LIST facility should be tight. Before the evaluation is begun,
the line should be tested with a leak detection system that has a known
performance. This protocol recommends that a tightness test be
performed on each pipeline system, because if one or more of the
pipelines is not tight, the performance of the system being evaluated will
be unnecessarily degraded.
Step 6 - Measure the pipeline compressibility characteristics. The
compressibility characteristics of the pipeline systems included in the
evaluation should be measured and reported. There is no minimum
specification to be met. Measurements of B/V0 and B should be made
when the temperature changes of the product in the line are small (i.e.,
less than 0.01°C over the duration of the measurement period) and
should follow the procedure given in Section 4.3. The leak detection
system should not be physically present in the line if it affects the
magnitude of B/V0. Unless temperature sensors such as thermistors are
used to measure temperature in the line, measurements of B/V0 and B
cannot be made until the pressure in the line stays within 1 psi over a
period equal to the average duration of a B/V0 measurement (i.e.,
approximately 2 min). Three estimates of B/V0will be made and the
median value reported.
79
-------
If the pipeline leak detector was removed or isolated from the line during
the compressibility tests, it should now be reconnected so that the leak
detection tests can be conducted.
The results of these measurements should be tabulated and reported on
Attachment 3 in Appendix B.
Step 7 - Determine the performance characteristics of the
instrumentation. Estimates must be made of (1) the minimum quantity
detectable by the system, (2) the precision and accuracy of each
instrument used to collect the data over the dynamic range of each
instrument required for the measurements, and (3) the response time of
the system. The resolution and flow rate of the threshold in gallons per
hour must also be reported. These measurements can be made in a
special calibration unit or on the pipeline system itself when the noise is
negligible. The general procedures required to estimate the performance
characteristics of the instrumentation are described in Section 6.1.
Step 8 - Develop a relationship between the leak and the output of
the measurement system. In Option 3 it is impractical to develop a
signal-plus-noise histogram at the EPA-specified leak using the direct
approach. This histogram is generated instead from the relationship
between the signal and the noise. This relationship must be verified by
means of experiments at one of the operational LIST facilities. The two-
step procedure for checking this relationship is described in Section 4.2.3.
The test results should be summarized in the tables in Attachment 7 in
Appendix B. The appropriate forms from Attachments 4 and 5, which
describe the temperature and leak conditions, as well as the test results,
should also be completed.
Steps 9 and 10 - Collect the noise data, the signal-plus-noise data,
and the trapped vapor data. Leak detection tests performed over a 6-
to 12-month period at each site follow the guidelines established in Step
3. This data collection procedure will yield an estimate of the noise
histogram that covers the temperature conditions under which the leak
detection system will actually be used. All leak detection tests should be
performed according to the manufacturer's protocol. The results of each
test should be recorded in terms of the output of the system. All leak
detection tests should begin immediately after dispensing operations
have ceased. This is important because the rate of change of the
temperature of the product in the pipeline decreases exponentially after
the last dispensing of product. Dispensing through other pipelines buried
80
-------
in the same backfill and in close proximity to the pipeline being tested
(i.e., within 12 in. of it) should also be terminated.
The results of these tests constitute the data needed to build the
histograms of the noise and the signal-plus-noise at the EPA-specified
leak rate. If a multiple-testing procedure is used, noise and signal-plus-
noise histograms must be compiled from the data used to determine
whether the pipeline is leaking and from the first test of the multiple-test
sequence; refer to Section 3.2.3 for additional details. The test results
should be partitioned into the following groups:
(1) for all pipeline systems and all operational LIST facilities: tests
that were started within 6 h of a delivery and within 30 min of the
last dispensing operation
(2) for each pipeline system at each operational LIST facility: tests
that were started within 6 hoi a delivery and within 30 min of the
last dispensing operation
(3) for each operational LIST facility where more than one pipeline
system was used: tests that were started within 6 hoi a delivery
and within 30 min of the last dispensing operation
(4) for all pipeline systems and all operational LIST facilities: tests
that were started within 12 h of a delivery and within 30 min of the
last dispensing operation
(5) for each pipeline system at each operational LIST facility: tests
that were started within 12 h of a delivery and within 30 min of the
last dispensing operation
(6) for each operational LIST facility where more than one pipeline
system was used: tests that were started within 12 hoi a delivery
and within 30 min of the last dispensing operation
Compute the mean, standard deviation, and 95% confidence intervals on
the means and standard deviations for the data in each of the data sets in
(1) through (6). The formulas necessary to perform these calculations are
given in Appendix E.
If at least two-thirds of the tests on each pipeline (i.e., at least 8 tests out
of 12) were started within 6 h of a delivery, the data in (1) should be used
to develop the noise histogram. Otherwise, the data from (4) should be
used. The signal-plus-noise histogram at the EPA-specified leak rate is
generated from the histogram and the relationship between the signal and
81
-------
the noise generated in Step 8. The relationship between the signal and
noise is used to shift the noise histogram appropriately. A performance
estimate is made from the PD/PFA analysis presented in Section 6.4. The
leak rate is defined at a line pressure of 20 psi, and the performance
estimate should be presented in those terms. If more than 25 tests are
available on any pipeline system in (2) or any operational LIST facility in
(3), additional performance estimates can be made. Estimates of
performance can also be made as a function of time after delivery, after
the last dispensing of product through the pipeline, or both, if data are
available. Such an analysis, while not part of this protocol, can be useful
in improving the performance of the leak detection system.
The temperature and leak conditions and the tests results obtained for
these conditions should be tabulated and reported on Attachments 4 and
5 in Appendix B.
Step 12 - Performance Analysis. The performance of the system can
be calculated from the data partitioned for specific leak rates, PDs and
PFAS. The protocol is designed so that the PD and PFA of the system are
established at the manufacturer's threshold and at the leak rate specified
by the EPA regulation (i.e., 0.1, 0.2, or 3 gal/h) at a test pressure of 20
psi. So that each system can be compared to others, a table for reporting
a variety of other performance estimates is provided as Attachment 2 in
Appendix B. If the leak detection system uses a multiple-test procedure,
performance estimates should follow the system's protocol, and
histograms should be generated from the data from both the last test and
the first test. The analysis of the performance of a detection system in
terms of PD and PFA is described in Section 6.4.
Step 13 - Evaluation Report. The results of the evaluation are tabulated
and reported in the standard format presented in Appendix A and
Appendix B. The performance characteristics of the instrumentation and
the performance estimates of the system's ability to detect leaks under
ambient environmental conditions will be presented in a standard set of
tables. The report should indicate whether the performance estimate was
made with data collected within 6 h or 12 h of a delivery of product. A
leak detection system meets the EPA standard for the leak rate specified
in the regulation if the calculated PD is 0.95 or greater and the PFA 0.05 or
less. The data used in the evaluation should be tabulated and included
as part of the evaluation report; this includes the date and time of the start
and end of each test, the test results, the time of the last dispensing
operation, and the date and time of the most recent delivery of product to
the tank. In addition, a general description of the pipeline systems used
82
-------
in the evaluation should be presented, including the operating pressure
and the bulk modulus of each pipeline system. Finally, a section is
provided for general comments.
6.3.4 Option 4 - Collect Data over a 6- to 12-month Period at 10 or More
Operational LIST Facilities
Option 4 is like Option 3 in that testing is conducted on a large number of
operational pipeline systems. It differs from Option 3 in that the integrity of the
pipelines may not be known. Otherwise, the two are identical. Option 4 includes
the same range of climatic conditions and requires the same number of tests per
pipeline as Option 3.
Like Option 3, it is best implemented when the relationship between the signal
and the noise is well known, and it is best suited to automatic systems that
routinely conduct a test of the pipeline whenever the LIST facility closes.
The histogram of the noise must be determined from analysis of the data. Since
the status of the lines is not known, it is possible that some of the test results
used to generate the histogram of the noise may be derived from lines with leaks.
If all data are used in the analysis, the procedure developed for Option 3 can be
followed directly. In some instances, it may be obvious that a line has a leak;
those data can be removed for the analysis if field investigation supports this
observation. However, removal of data from the analysis should be done with
extreme care and should be clearly explained in the evaluation report. Removing
data from the analysis is not justified, for example, simply because the test
results from one pipeline (or a few test results from one or more pipelines) are
significantly different from the majority of the test results. Any removal of data
can bias the results, i.e., increase performance. Therefore, data should be
removed only if it has actually been determined, through a special field test, that
the line is leaking, or if it can be shown that anomalous results are due to
instrumentation or equipment problems. In some cases, special data analysis
strategies can be developed to statistically separate the test results derived from
lines believed to be leaking from results derived from lines that are not leaking.
The histogram of all the data and the histogram of the data actually used to
develop the noise histogram should both be presented if any data have been
removed. This approach will normally provide the largest database with which to
make an evaluation but also requires the most care to characterize the histogram
of the noise.
83
-------
6.3.5 Option 5 - Develop the Noise and Signal-plus-noise Data from an
Experimentally Validated Computer Simulation
In Option 5, models of the important sources of noise that control the
performance of the leak detection system are developed and validated through a
comprehensive set of experiments. These models are then used with models of
the leak detection system to simulate the performance of the system over a wide
range of conditions. The simulation results must then be checked
experimentally. This check requires a set of tests with the actual leak: detection
system. In some cases, models of the noise cannot be developed with sufficient
accuracy for the evaluation. In such a case, a database of measured conditions
is collected and is used instead of the model. In general, Option 5 should be
used with caution, because it is more difficult to implement properly than the
other, more direct options for evaluating performance.
There are, however, a number of advantages to Option 5. First, this option is
particularly useful if many systems of the same type are to be evaluated and
compared, because each system will be tested under identical conditions.
Second, with this option the performance estinlates can be extended over a
wider range of leak rates and pipeline configurations. Third, it is possible to limit
the number of actual field tests with the leak detection system.
There are also a number of disadvantages to Option 5. First, it takes a
significant technical effort to identify and develop simulation models of the
sources of noise, and these models are necessary before any leak detection
system can be evaluated. Second, accurate performance estimates require that
all sources of noise that affect the particular leak detection system being
evaluated be identified and included in the simulation. Third, accurate
performance estimates require that each source of noise be properly modeled.
Fourth, the operational practice, particularly the influence of the test operator, is
usually not included in the evaluation, yet it may have a significant impact on
performance. Fifth, for a number of reasons, it is easier to misuse Option 5 than
it is the first four, especially because the evaluation conditions will ultimately
become known and leak detection systems will be designed to perform well
under these known conditions.
The disadvantages of computer simulation as an evaluation approach actually
emphasize the strength of simulation as a design tool. Because tradeoffs in
performance under a wide range of conditions can readily be examined,
simulation is put to better use when it is applied to designing the specifications of
a leak detection system rather than to evaluating its performance. Needless to
say, if a simulation is used to develop a system, the same simulation should not
also be used to evaluate it.
84
-------
Option 5 can be used only when a leak detection system can be accurately
described mathematically, when the models of the noise are validated
experimentally, and when the simulation results are verified by means of
experiments conducted with the actual system. Option 5 was the approach used
to evaluate the performance of volumetric tank tightness test methods in the EPA
program on that subject [7].
Only a general outline of the steps in Option 5 is provided below. This is
because a different set of noise, signal-plus-noise, and leak detection system
models would be required for each type of system to be evaluated. (For a
description of the statistical topics discussed below, see Appendix E.)
Step 1. Develop a probability distribution, P(N), for any noise source other than
temperature that is applicable to the system being evaluated.
(Temperature effects are included in Step 3 below.) The P(N) may be
derived empirically from the data or may be derived from a
mathematical model that has been developed, validated
experimentally, and exercised over a full range of conditions.
Step 2. Develop and validate experimentally a relationship between the signal
and the noise.
Step 3. Develop a computer model of the leak detection system. The model
should include:
• all quantities that are measured by the system
• the resolution, precision, accuracy, and dynamic range of the
system's sensors
• any waiting periods that are included in the test protocol
• any deliveries and/or dispensing included in the test protocol
• the test duration as defined by the system's test protocol
• the data sampling rate
• the data analysis procedure
• the detection criterion
In addition, the output of the model must be in units of flow rate, and so
conversion routines should be included in the model as needed.
Step 4. Develop the test simulation using
85
-------
a) a heat-transfer model or a comprehensive set of field data to
determine the rate of change of product temperature in the pipeline for
a given set of ground and tank temperature conditions and a given set
of dispensing conditions,
b) a model to estimate how the rate of change of product temperature
affects any of the other noise sources and the output quantity being
measured,
c) the relationship between height and volume in a given container to
obtain the rate of change of volume for the noise sources defined in
Stepl,
d) development of the signal-plus-noise probability distribution from
Steps 2 and 4 (c), and
e) the leak detection system model developed in Step 2 in conjunction
with pressure/volume/temperature relationships to determine the test
outcome for a specified leak rate and the conditions described by
Step 4 (a), (b), and (c)
Step 5. Validate the simulation with data obtained from a minimum of five
actual leak detection tests on a nonleaking pipeline and five on a
pipeline leaking at a known rate. The leak rate generated for the five
leaking-pipeline tests should be equal to the leak rate at which the
performance estimate will be made.
For all ten tests, the noise sources should be controlled, i.e., set to
specific values which can then be used as input to the simulation. If all
ten tests are within 15% of the results obtained by the simulation, the
simulation is considered valid. The nominal temperature differences
between the ground and the product dispensed through the pipeline
system for an hour should be approximately -15, -7.5, 0, +7.5, and
Step 6. Follow the steps in Option 1 to complete the evaluation, with one
exception. Instead of conducting the field tests in Step 1 1 of Option 1 ,
use the simulation to derive the data required to develop the noise and
the signal-plus-noise histograms. The simulation should be exercised
under the same conditions required by Option 1; all other field
measurements, such as the measurement of the performance
characteristics of the instrumentation, should be made in the same way
as in Option 1. The tests required to estimate the sensitivity of the
system to trapped vapor are usually done experimentally; they can be
86
-------
simulated if trapped vapor is one of the sources of noise included in
the computer model.
6.4 CALCULATION OF PD AND PFA
The steps for calculating the PFA and the PD at a leak rate, LR, are given below,
along with an example of how these calculations are done. These sample
calculations are for tests conducted under the same temperature conditions on a
nonleaking pipeline and for tests on a pipeline with a leak of 0.1 gal/h defined at 20
psi. The data collected on the nonleaking pipeline are used to generate a cumulative
frequency distribution of the noise, and the data collected on the leaking pipeline are
used to generate a cumulative frequency distribution of the signal-plus-noise. The
same analysis procedure can be used if the cumulative frequency distribution of the
signal-plus-noise is generated from an experimentally validated relationship between
the signal and the noise and the cumulative frequency distribution of the noise. An
example of how to estimate the probability of detection from this approach is also
given. (See Tables 6.1 and 6.2.) In this example, it is assumed that the signal is
independent of the noise and simply adds with the noise. The estimates of the PFA
and the PD at a leak rate, LR, are for a specific threshold, T.
Estimating the probability of false alarm is done as follows.
1. Tabulate the available results of tests performed on a nonleaking pipeline,
arranging them in order from the lowest value to the highest and numbering
them sequentially (1 being the lowest).
2. Assign an individual frequency to each test result equal to 1/(n + 1), where n
is the total number of test results.
3. Develop the cumulative frequency for each test result by multiplying the
individual frequency of each result by the number assigned to each test in
Step 1. The results are shown in Table 6.1. For example, the fifth test result
would have a cumulative frequency of 0.192, which is equal to 5 times the
individual frequency (i.e., 5/(n + 1)), and a flow rate of -0.031 gal/h.
4. Generate a curve by plotting the test result on the abscissa (x axis) versus
the cumulative frequency on the ordinate (y axis). This curve is the
cumulative frequency distribution of the noise (the distribution of test results
from nonleaking tanks), and corresponds to Figure 2.2.
87
-------
Table 6.1. Values of the Cumulative Frequency Distribution of the Noise Shown in Figure 2.2
Cumulative
Frequency
0.038
0.077
0.115
0.154
0.192
0.231
0.269
0.308
0.346
0.385
0.423
0.462
0.500
Test Result
(gal/h)
-0.092
-0.052
-0.042
-0.037
-0.031
-0.025
-0.015
-0.011
-0.010
-0.007
-0.005
-0.004
-0.002
5. Locate the threshold on the abscissa
6. Estimate the
PFA from the intersection
Cumulative Test Result
Frequency
0.538
0.577
0.615
0.654
0.692
0.731
0.769
0.808
0.846
0.885
0.923
0.962
of the curve generated in
of the threshold and the
(gal/h)
0.000
0.003
0.008
0.009
0.014
0.020
0.022
0.023
0.027
0.031
0.042
0.056
Step 4.
cumulative
distribution curve. This value is read from the ordinate at the intersection
point. For a threshold of -0.05 gal/h, the PFA equals 0.085 for the data plotted
in Figure 2.4. This value can also be estimated by interpolation of the data in
Table 6.1.
The PFA can also be estimated from an analysis of how often the threshold
was exceeded. The PFA is calculated by dividing the number of times the
threshold was exceeded by the total number of tests plus one. For the noise
data in Table 6.1, PFA = 21(25 + 1) = 0.077.
Estimating the probability of detection at a specified leak rate (where the
cumulative frequency distribution of the signal-plus-noise is generated from
data collected on a pipeline with a specific leak) is done as follows.
1. Tabulate the available results of tests performed at the leak rate of interest,
arranging them in order from the lowest value to the highest and numbering
them sequentially (1 being the lowest).
2. Assign an individual frequency to each test result equal to 1/(n + 1), where n
is the total number of test results.
3. Develop the cumulative frequency for each test result by multiplying the
individual frequency of each result by the number assigned to each test (Step
1). The results are shown in Table 6.2. For example, the fifth test result
would have a cumulative frequency of 0.192, which is equal to 5 times the
individual frequency (i.e., 5/(n + 1)), and a flow rate of-0.131 gal/h.
88
-------
Table 6.2. Values of the Cumulative Frequency Distribution of the Signal-plus-noise Shown
in Figure 2.3 Generated for Leak Rate (i.e., Signal) of 0.10 gal/h
Cumulative
Frequency
0.038
0.077
0.115
0.154
0.192
0.231
0.269
0.308
0.346
0.385
0.423
0.462
0.500
Test Result
(gal/h)
-0.192
-0.152
-0.142
-0.137
-0.131
-0.125
-0.115
-0.111
-0.110
-0.107
-0.105
-0.104
-0.102
Cumulative
Frequency
0.538
0.577
0.615
0.654
0.692
0.731
0.769
0.808
0.846
0.885
0.923
0.962
Test Result
(gal/h)
-0.100
-0.097
-0.092
-0.091
-0.086
-0.080
-0.078
-0.077
-0.073
-0.069
-0.058
-0.044
4. Generate a curve by plotting the test result on the abscissa (x axis) versus
the cumulative frequency on the ordinate (y axis). This curve is the
cumulative frequency distribution of the signal-plus-noise (the distribution of
test results from a pipeline with a leak of 0.1 gal/h). Negative values mean
that product is flowing out of the tank or pipeline. This curve corresponds to
Figure 2.3.
5. Locate the threshold on the abscissa of the curve generated in Step 4 under
"Estimating the probability of false alarm."
6. Estimate the PD from the intersection of the threshold and the cumulative
frequency distribution curve. This value is read from the ordinate at the
intersection point. For a threshold of -0.05 gal/h, the PD equals 0.945 for the
data plotted in Figure 2.4. This value can also be estimated by interpolation
of the data in Table 6.1. Other estimates of PD can be made against a
particular leak rate by changing the threshold.
The PD can also be estimated from an analysis of how often threshold was
exceeded for a particular leak rate. The number of times the threshold was
exceeded is divided by the total number of tests plus one. For the signal-
plus-noise data in Table 6.2, PD = 24/(25+1) = 0.923.
7. Other estimates of PD can be made as a function of threshold and leak rate if
the signal-plus-noise data have been collected for that leak rate or if the
relationship between the signal and the noise can be developed from the
existing cumulative frequency distributions. For each new leak rate (signal-
89
-------
plus-noise) curve, the effects on the PD of changing the threshold can be
estimated directly from the intersection of the threshold with the curve.
Estimating the probability of detection at a specified leak rate (where the
cumulative frequency distribution of the signal-plus-noise is generated from
the noise cumulative frequency distribution and an experimentally validated
relationship between the signal and the noise) is done as follows.
1. Generate a cumulative frequency distribution of the signal-plus-noise for a
specific leak rate, LR, by adding the system's response to the leak to each
data point included in the cumulative frequency distribution of the noise using
the manufacturer's relationship between the signal and the noise. If, for
example, the signal is simply additive with the noise, the signal-plus-noise
cumulative frequency distribution for an outflowing leak rate of 0.10 gal/h is
obtained by adding -0.10 gal/h to each of the tabulated test results generated
in Step 4 (i.e., in Table 6.1 ). This results in a shift of-0.10.g/h in the
cumulative frequency distribution of the noise.
2. Proceed with Steps 4 through 7 above.
90
-------
SECTION 7
EVALUATION PROCEDURE FOR SYSTEMS
THAT USE A PRESET THRESHOLD
Some leak detection systems do not report the output quantity. Instead, they are
designed to respond only if the output quantity is large enough to activate a preset-
threshold switch. The procedure for evaluating preset-threshold systems differs only
slightly from that for systems which report an output quantity. Many of the leak detection
systems designed to meet the 3-gal/h hourly test requirement established in the EPA
regulation use a preset threshold.
7.1 PERFORMANCE CHARACTERISTICS OF THE INSTRUMENTATION
Before performing any evaluation experiments with a preset-threshold leak detection
system, it is necessary to ensure that the system is working correctly and will
respond when the preset threshold is exceeded. This can be done with a simple
calibration procedure. Depending on which option is selected, these measurements
may or may not be used quantitatively in the calculation of the PD and PFA of the
system, but they are used qualitatively to determine the advisability of proceeding
with the evaluation. If the instrumentation (or system) noise is so large that the
required performance could not be achieved even if no other noise sources were
present, the evaluation procedure could be stopped, and a reassessment of the
system design might be considered.
The calibration should consist of a sequence of measurements with the leak
detection system in a controlled environment to determine what the system is
measuring (i.e., specificity) and to make an estimate of the resolution, precision,
accuracy, and minimum detectable signal. Preset-threshold systems, like those that
report a flow rate, do measure a physical quantity, which is what triggers the
threshold switch. The difference is that this quantity is not reported. For most
systems that measure a physical quantity (for example, volume or pressure), the
specificity is obvious. The resolution of the system is the smallest division for which
a quantity is measured; since the resolution is usually well known, it does not have to
be measured as part of this protocol, but it does have to be reported. The minimum
detectable quantity is defined in this protocol as that quantity that can be detected
with a PD of 0.95 and a PFA of 0.05; assuming that the instrumentation noise is
normally distributed, the minimum detectable signal is 3.3 times larger than the
precision.
The flow rate at which the threshold of the measurement system is exceeded, as well
as the precision and accuracy of system, can be determined from the tests described
91
-------
below. These tests should be done on a pipeline system in which the temperature
changes are negligible. The procedure is as follows:
• Determine the threshold. An estimate of the flow rate at which the
threshold will be exceeded and at which the system will signal the presence
of a leak is required. This flow rate is defined at a pressure of either 10 psi
(for 3-gal/h hourly testing systems) or 20 psi (for all other systems). For leak
rates of 0.1 gal/h (for a line tightness test) and 0.2 gal/h (for a monthly
monitoring test), the flow rate at which the threshold will be exceeded should
be measured to within 0.015 and 0.030 gal/h, respectively; for a leak rate of
3.0 gal/h (for an hourly test), the flow rate at which the threshold will be
exceeded should be measured to within 0.25 gal/h. The sources of ambient
noise in the pressurized pipeline system that will be used in the evaluation
should be minimized. Different leak rates are generated, from small to large,
until the threshold is exceeded.
• Determine the minimum detectable signal. The minimum detectable
signal is less than or equal to the threshold.
• Determine the precision. The leak rate at which the threshold is exceeded
is found by repeating the leak detection test a number of times, with the
difference in the size of each leak rate getting progressively smaller until the
system responds. The precision of the system is determined from the
standard deviation of the five flow rates at which the threshold was exceeded
is the precision of the system. The uncertainty of the precision estimate
made with this method is dependent on the size of the increment between
leak rates; as fine an increment as possible should be used.
• Determining the accuracy. The accuracy of the system is determined from
the mean of the five flow rates used to estimate precision. The accuracy is
the difference between the measured flow rate and the flow rate at which the
manufacturer claimed that the system would respond. If no claim is made, an
accuracy measurement cannot be calculated or reported.
7.2 DEVELOPMENT OF THE NOISE AND THE SIGNAL-PLUS-NOISE DATA
In this protocol, it is assumed that the evaluation is being performed to obtain the PD
and PFA at the leak rate specified in the EPA regulation for the type of system being
evaluated, e.g., 0.1 gal/h for a line tightness test, 0.2 gal/h for a monthly monitoring
test, or 3 gal/h for an hourly test. Thus, the procedure described below leads to the
development of the noise and the signal-plus-noise data for the leak rate of greatest
regulatory interest for a line tightness test, a monthly monitoring test, and an hourly
92
-------
test. If local regulations specify leak rates more stringent than those in the EPA
regulation, the local standard can be substituted for the EPA-specified leak rates.
Unlike those leak detection systems that quantitatively measure and report the
output of the system, the only output from a preset-threshold system is a simple pass
or fail* - i.e., whether or not the system responded to the leak or the temperature
condition. As a consequence, this is the only performance estimate that can be
derived from the evaluation. It is not possible to examine the tradeoffs in
performance by changing the threshold. An advantage of preset-threshold systems
is that the analysis used to estimate PFA and the PD for the EPA-specified leak rate is
simpler than it is for the systems that quantitatively measure the output; however, the
latter can be analyzed the same way as the preset-threshold systems. The method
of analysis is described in Section 7.4.
If the system uses a multiple-test strategy, the histogram of the noise and the
histogram of the signal-plus-noise are generated from that test sequence which was
the basis for declaring a leak. In addition to histograms used to develop a
performance estimate of the system, a second performance estimate is requested.
This second estimate is based only on the results of the first test in the multiple-test
sequence. Refer to Section 3.2.3 for a discussion of multiple-test strategies.
7.3 EVALUATION PROCEDURE
The same five options for estimating the performance of the leak detection systems
that report an output quantity are used to collect the data necessary to characterize
the noise and the signal-plus-noise for systems that use a preset threshold. These
options are presented in Section 6.3 and are not repeated here. There are only a
few minor differences. First, the performance characteristics are determined
according to the procedures presented in Section 7.1 and not Section 6.1. Second,
the analysis required to estimate performance in terms of PD and PFA follows the
procedures presented in Section 7.4 and not Section 6.4. Third, the noise and the
signal-plus-noise histograms must be measured directly.
Some systems that use a preset-threshold switch and are intended to meet the 3-
gal/h hourly test requirements are designed to do a quick test of the pipeline system.
Normally, the duration of a test ranges from a few seconds to tens of seconds
because the system is designed to test the line at least once per hour between
occurrences of product dispensing. Whereas most other systems have a test
duration equal to the data collection time (i.e., the data that will be used in calculating
a flow rate that will be compared to a threshold), the systems in question have a test
Pass means that the threshold was not exceeded and fail means that the threshold was
exceeded.
93
-------
duration equal to the difference between the time a system is activated and the time
it responds to a leak. In these systems, the test duration may not be specifically
defined, since the system does not control the response time. To avoid misleading
or ambiguous results with these systems, therefore, the evaluator should ensure that
the test duration is clearly defined in the manufacturer's test protocol. For the
purposes of the evaluation, a test duration must be specified. The duration should
be consistent with the normal operational practice and the manufacturer's intended
use of the system. If it is not, the evaluator should clearly point this out in the report,
for it may mean that the system being evaluated is not the same as the system being
sold commercially in the sense that the system may not respond as quickly as (i.e.,
may have a longer test duration than) the user expects.
7.4 CALCULATION OF PD AND PFA
The performance analysis is done as follows. The PFA is determined directly from
the number of times the threshold was exceeded (the number of times the pipeline
failed the test) in the zero-leak-rate data (the noise data) divided by the total number
of tests plus one. Estimates of PD can be made directly from the tests conducted at
the EPA-specified leak rate and any other leak rate for which adequate data are
available (i.e., 25 tests over the full range of temperature conditions). The PD is the
number of times the threshold was exceeded divided by the total number of tests
plus one. The analysis is a simple tabulation. The data in Tables 7.1 and 7.2 are the
same data found in Tables 6.1 and 6.2, but they are reproduced as if they had been
collected with a preset-threshold leak detection system instead of one that reports an
output quantity.
Table 7.1. Values of the Cumulative Frequency Distribution of the Noise Shown in Figure 2.2
Cumulative
Frequency
0.038
0.077
0.115
0.154
0.192
0.231
0.269
0.308
0.346
0.385
0.423
0.462
0.500
Test Result
(gal/h)
Fail
Fail
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Cumulative
Frequency
0.538
0.577
0.615
0.654
0.692
0.731
0.769
0.808
0.846
0.885
0.923
0.962
Test Result
(gal/h)
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
Pass
94
-------
Table 7.2. Values of the Cumulative Frequency Distribution of the Signal-plus-noise Shown
in Figure 2.3 Generated for Leak Rate (i.e., Signal) of 0.10 gal/h
Cumulative
Frequency
0.038
0.077
0.115
0.154
0.192
0.231
0.269
0.308
0.346
0.385
0.423
0.462
0.500
Test Result
(gal/h)
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Cumulative
Frequency
0.538
0.577
0.615
0.654
0.692
0.731
0.769
0.808
0.846
0.885
0.923
0.962
Test Result
(gal/h)
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Fail
Pass
The test results given in Tables 7.1 and 7.2 are derived from a system having a
threshold switch set to -0.05 gal/h and subject to the same conditions as the system
that reports an output quantity (see Tables 6.1 and 6.2). When the probability of
false alarm is calculated from the test results in Table 7.1, PFA = 27(25+1) = 0.077.
When the probability of detection against a leak rate of 0.1 gal/h is calculated from
the test results in Table 7.2, PD= 247(25+1) = 0.923. (It should be noted that in an
actual test, the data will not already have been sorted as has been done for the data
in Tables 7.1 and 7.2.)
95
-------
SECTION 8
LEAK DETECTION TESTS
WITH TRAPPED VAPOR IN THE PIPELINE
If Option 1, 2, or 5 is used to characterize the noise and signal-plus-noise histograms, a
special set of three tests will be conducted with a small volume of vapor trapped in the
pipeline. These tests are intended to determine the sensitivity of the leak detection
system to any residual vapor that might be trapped in a line during a test. The results of
these three tests will be tabulated and reported, but will not be included in the histogram
of the noise or signal-plus-noise used to estimate the performance of the system.
Trapped vapor tests are not required in Options 3 and 4 because these options both
require many tests at a large number of operational LIST facilities; as a result, it is likely
that trapped vapor will be present during some of the tests and that it will thus be
included in the actual performance estimates.
If the system is being evaluated as a line tightness test or a monthly monitoring test, the
three tests will be conducted with leaks of 0.0, 0.1, and 0.2 gal/h, and with vapor trapped
in the pipeline. The amount of trapped vapor will be that produced by a 6.4-in.3 ± 0.6 in.3
(105-ml ± 10 ml) vapor pocket apparatus. These tests should be done under the same
nominal temperature condition. If these are blind tests, the tests will be randomly mixed
in with the other tests in the test matrix used to develop the noise and the signal-plus-
noise histograms. If the system is being evaluated as an hourly test, the leaks
generated for the three tests should be 0, 2.75, and 3.25 gal/h, respectively. If these are
blind tests, the leaks should be in random order.
The vapor pocket apparatus shown in Figure 4.7 on page 49, which has been specially
designed for this protocol, can be used to trap vapor in the pipeline. Trapped vapor is
introduced in the line by opening or closing an inlet valve. Section 4.5.1 describes the
apparatus and how it can be used to generate a vapor pocket.
The results of these three tests will be reported in Attachment 6 in Appendix B.
96
-------
SECTION 9
REPORTING OF RESULTS
A form on which to summarize the results of the evaluation has been provided in
Appendix A. The form requires that the following information be provided:
• the name of the leak detection system that was evaluated and the name and
address of its manufacturer
• the performance of the system for detection of a leak equal to the one
specified in the EPA regulation in terms of probability of detection and
probability of false alarm
• the criterion for declaring a leak, including (1) whether the system is one that
reports the output and compares it to a threshold or whether it is one that
uses a preset threshold, (2) the flow rate of a leak represented by the
threshold, and (3) whether the system uses a multiple-test strategy
• the option used to collect the data for the evaluation
• a brief description of the pipeline system(s) used in the evaluation
• a summary of the range of temperature conditions used in the evaluation
• a summary of the leak rates used to make the performance estimate
• a summary of the sensitivity of the system to the presence of trapped vapor in
the pipeline
• the performance characteristics of the instrumentation that comprises the
leak detection system
• a brief description of the types of pipeline systems to which the leak detection
system is applicable
• the important features of the protocol for conducting a test with this leak
detection system
• a list of attachments to the form
• the name, address and telephone number of the organization that conducted
the evaluation and the name, date, and signature of the individual who
certifies that the system was evaluated according to the procedures outlined
by the EPA
97
-------
There are seven attachments to the form that give additional details about the system
and the evaluation. With the data and information provided in these attachments, all of
the results of the evaluation could be independently reviewed and verified. The seven
attachments include:
• Attachment 1 - Description of the System Evaluated
• Attachment 2 - Summary of the Performance of the System Evaluated
• Attachment 3 - Summary of the Configuration of the Pipeline System(s) Used
in the Evaluation
• Attachment 4 - Data Sheet Summarizing the Product Temperature Conditions
Used in the Evaluation
• Attachment 5 - Data Sheet Summarizing the Test Results and the Leak Rates
Used in the Evaluation
• Attachment 6 - Data Sheet Summarizing the Test Results and the Trapped
Vapor Tests
• Attachment 7 - Data Sheet Summarizing the Test Results Used to Check the
Relationship Supplied by the Manufacturer for Combining the Signal and the
Noise
98
-------
SECTION 10
TECHNICAL BASIS FOR VALUES USED IN THE PROTOCOL
The technical basis for the choice and number of test conditions is discussed below.
10.1 RANGE OF TEMPERATURE CONDITIONS
The range of temperature conditions generated for an evaluation is based on a study
completed for the EPA [6,7]. The study estimated the average monthly difference in
temperature between the air and the product in the tank. It can be assumed that the
average air temperature is approximately equal to the temperature of the ground to a
depth of 1 to 3 ft. The temperature of the product brought into the pipeline was
estimated from empirical measurements made in underground storage tanks. Data
from 77 cities throughout the United States were used to generate a histogram of
these average differences. These data were collected during the two months that
had the coldest and hottest average temperatures, i.e., January and July,
respectively. The shapes of the histograms were nearly identical, i.e., the standard
deviations were approximately equal, but the means were different. The study
indicated that the mean temperature differences during January and July were -27°F
(-15°C) and +9°F (+5°C), respectively. The standard deviation of the temperature
differences for each month was approximately 9°F. If it is assumed that there is a
similar distribution for each of the two months and a mean that is uniformly
distributed between the minimum and maximum values determined by the January
and July means, the temperature is approximately normally distributed. The
temperature conditions selected for this protocol and shown in Table 5.1 are based
on this analysis.
10.2 NUMBER OF TESTS
The number of independent tests required to evaluate the performance of a pipeline
leak detector depends on the statistical uncertainty desired for the PD and PFA.
Independence means that the individual tests are not correlated with each other. A
high degree of correlation is found if the testing errors are systematic rather than
random. When this is the case, the same error occurs in each individual test and the
averaging effect, which can reduce the noise fluctuations, is not realized. If the tests
are not independent, a larger number of tests is required if the same uncertainty is to
be maintained. Most pipeline testing errors tend to be systematic, and a high degree
of correlation is generally found, as, for example, when successive tests are
conducted over a short time during which there are no temperature changes in the
line, or when trapped vapor is present during each test. Since changes in the
temperature of the product are the largest source of error in a vapor-free pipeline
system, independence will be achieved if a different temperature condition is created
99
-------
for each test. A new temperature condition can be generated by pumping in product
whose temperature is different from that of the product in the pipeline and that of the
surrounding ground. At a test facility, a full range of temperature conditions can be
created over a short-period of time (two to four weeks) if the product can he heated
or cooled before it is transferred to the line. If the tests are done at an operational
LIST facility, a new temperature condition is created each time there is a new
delivery of product to the tank system. However, consecutive deliveries do not
necessarily produce independent temperature conditions, because over a period of
several weeks the temperature of the product delivered to the tank system and that
of the ground surrounding the system tend to be similar. To guarantee a wide range
of temperature conditions, data must be collected over a 6- to 12-month period. In
order to avoid biasing the performance toward either the high or low end of the scale,
the data from the LIST facilities must be partitioned into groups according to the
number of hours that have elapsed after a product delivery.
An estimate of the number of independent tests was made; it was assumed that the
95% lower and upper confidence intervals on the PD and PFA, respectively, gave a PD
no lower than 0.90 and a PFA no higher than 0.10. This means that there is a
probability of 95% that an instrument that has a PD of 0.95 and a PFA of 0.05 would
have experimental PD/PFA values greater than 0.90 and 0.10, respectively. The
estimate assumes that the cumulative frequency distribution (CFD) of the noise and
the signal-plus-noise are normally distributed and that a threshold consistent with a
PFA of 0.05, the EPA minimum requirement, is used. It is further assumed that the
signal is independent and additive with the noise. This means that the signal-plus-
noise CFD is simply a shifted replica of the noise, i.e., the mean is equal to the signal
and the standard deviation is the same. For this performance model, the PFA and PD
can be determined from the standard deviation of the noise and signal-plus-noise
CFDs. If it is assumed that the mean of the noise CFD is zero (i.e., that it has no
bias), the 95% confidence interval on the standard deviation of the histograms is
determined by the x2 (chi-squared) probability distribution. The 95% confidence
intervals are determined by the number of independent tests. The uncertainty is
large if the number of tests is small; the uncertainty decreases as the number of tests
increases.
If the normal probability density performance model is used, the leak rate that can be
detected with a PD of 0.95 and a PFA of 0.05 is equal to 3.28 standard deviations, and
the threshold is equal to 1.64 standard deviations. If the leak rate is 0.1 gal/h, the
standard deviation must be 0.03 gal/h; if the leak rate is 0.2 gal/h, the standard
deviation must be 0.06 gal/h. The corresponding thresholds are 0.05 and 0.10 gal/h,
respectively. If these thresholds are used, standard deviations of 0.039 for the 0.1-
gal/h leak rate and 0.078 gal/h for the 0.2-gal/h leak rate would result in a PFA of 0.10
and a PD of 0.90. Thus, the upper 95% confidence interval on a PFA of 0.05 and the
100
-------
lower 95% confidence interval on a PD of 0.95 would result in the detection of leak
rates of 0.128 and 0.258 gal/h, respectively, for the two leak rates of interest. These
calculations suggest that a minimum of 32 tests is required.
It was decided to select 25 as the minimum number of independent tests required for
the evaluation. (Statistically, the difference between 32 and 25 is very small.) The
value of the standard deviation, the minimum detectable leak rate, and the PD and
PFA for 25 independent tests defined by the 95% confidence intervals are
summarized in Tables 10.1 through 10.3 for the detection of leaks of 0.1, 0.2, and
3.0 gal/h with a PD of 0.95 and a PFA of 0.05. These confidence intervals suggest the
degree of uncertainty in estimating performance with 25 tests. Any experimental
leak rate-value determined from a 25-test evaluation that falls within the 95%
confidence intervals on the minimum detectable leak rates given in Tables 10.1
through 10.3 for a PD of 0.95 and a 0.05, or any PD and PFA that falls within the 95%
confidence intervals of the PDs and PFAS given in Tables 10.1 through 10.3, is not
statistically distinguishable from the 0.1-, 0.2-, and 3.0-gal/h EPA standards.
Table 10.1. Experimental Uncertainty on the Standard Deviation of the Noise and Signal-
plus-noise Histograms, the Smallest Leak Rates That Can Be Detected with a PD of 0.95 and
a PFA of 0.05, and the PD and PFA Characterized by the 95% Confidence Intervals on the
Standard Deviation for Detection of a Leak Rate of 0.10 gal/h
Quantity
Standard Deviation - gal/h
Smallest Detectable
Leak Rate - gal/h
PD
PFA
Lower Confidence
Interval
0.025
0.083
0.890
0.024
Mean
0.03
0.10
0.95
0.05
Upper Confidence
Interval
0.041
0.134
0.976
0.110
Table 10.2. Experimental Uncertainty on the Standard Deviation of the Noise and Signal-
plus-noise Histograms, the Smallest Leak Rates That Can Be Detected with a PD of 0.95 and
a PFA of 0.05, and the PD and PFA Characterized by the 95% Confidence Intervals on the
Standard Deviation for Detection of a Leak of 0.20 gal/h
Quantity
Standard Deviation - gal/h
Smallest Detectable
Leak Rate - gal/h
PD
PFA
Lower Confidence
Interval
0.050
0.166
0.890
0.024
Mean
0.06
0.20
0.95
0.05
Upper Confidence
Interval
0.0815
0.268
0.976
0.110
101
-------
Table 10.3. Experimental Uncertainty on the Standard Deviation of the Noise and Signal-
plus-noise Histograms, the Smallest Leak Rates That Can Be Detected with a PD of 0.95 and
a PFA of 0.05, and the PD and PFA Characterized by the 95% Confidence Intervals on the
Standard Deviation for Detection of a Leak of 3.0 gal/h
Quantity
Standard Deviation - gal/h
Smallest Detectable
Leak Rate - gal/h
PD
PFA
Lower Confidence
Interval
0.76
2.5
0.890
0.024
Mean
0.91
3.0
0.95
0.05
Upper Confidence
Interval
1.22
4.0
0.976
0.110
10.3 RANGE OF THE BULK MODULUS
The range of the bulk modulus (elasticity) is not well known for the population of
underground storage tank pipeline systems found throughout the United States.
Only several values of B have been measured. The value of B used in this protocol
is based on a limited set of data collected during a program conducted for the
American Petroleum Institute [4,5].
10.4 VAPOR POCKETS
Vapor trapped in the line can affect the performance of a leak detection system.
There are two effects. First, the trapped vapor changes the bulk modulus of the
pipeline system. This affects the magnitude of the conversion factor needed, for
example, to convert a pressure measurement to a flow rate. Second, if there is a
large amount of trapped vapor, thermally induced volume changes can affect the
performance of the system because volume changes also affect pressure changes in
the line. Some systems are particularly sensitive to the presence of trapped vapor
and others are not. According to the evaluation protocol, the pipeline system should
be as free of trapped vapor as possible. Thus, in general, the effects of trapped
vapor will not be included in the performance estimates. If the effects of trapped
vapor were included, the number of test conditions would have to be increased
significantly. Because trapped vapor can have a measurable impact on
performance, however, several tests must be done so that the sensitivity of the
system to trapped vapor can be determined.
102
-------
REFERENCES
1. U.S. Environmental Protection Agency, Underground Storage Tanks;
Technical Requirements, 40 CFR Part 280, Federal Register. Vol. 53. No.
185 (23 September 1988).
2. United States Environmental Protection Agency, Standard Test Procedures
for Evaluating Leak Detection Methods: Liquid-phase Out-of-tank Product
Detectors, EPA Contract No. 68-03-3409, Work Assignment 02 (March
1990).
3. United States Environmental Protection Agency, Standard Test Procedures
for Evaluating Leak Detection Methods: Vapor-phase Out-of-tank Product
Detectors, Technical Report, EPA Contract No. 68-03-3409, Work
Assignment 02 (March 1990).
4. J. W. Maresca, Jr., N. L. Chang, Jr., and P. J. Gleckler, A Leak Detection
Performance Evaluation of Automatic Tank Gauging Systems and Product
Line Leak Detectors at Retail Stations, Final Report, prepared for the
American Petroleum Institute, Vista Research Project 2022, Vista Research,
Inc., Mountain View, California (4 January 1988).
5. J. W. Maresca, Jr., M. P. MacArthur, A. M. Regalia, J. W. Starr, C. P. Wilson,
R. M. Smedfjeld, J. S. Farlow , and A. N. Tafuri, Pressure and Temperature
Fluctuations in Underground Storage Tank Pipelines Containing Gasoline,
Journal of Oil and Chemical Pollution (in press).
6. J. W. Starr and J. W. Maresca, Jr. Protocol for Evaluating Volumetric Leak
Detection Methods for Underground Storage Tanks, Technical Report,
contract No. 68-03-3255, Enviresponse, Inc., Livingston, New Jersey, and
Vista Research. Inc., Palo Alto, California (June 1986).
7. U.S. Environmental Protection Agency, Evaluation of Volumetric Leak
Detection Methods for Underground Fuel Storage Tanks, Vol. I (EPA/600/2-
88/068a) and Vol. II (EPA/600/2-88/068b) Risk Reduction Engineering
Laboratory, Edison, New Jersey (December 1988).
8. ASTM Subcommittee D.34.11 on Underground Storage Tanks, ASTM
Practice for Evaluating and Reporting the Performance of Release Detection
Systems for Underground Storage Tanks, Fifth Draft (17 August 1988).
103
-------
APPENDIX A
FORM TO PRESENT A DESCRIPTION
OF THE PIPELINE LEAK DETECTION SYSTEM
EVALUATED ACCORDING TO THE EPA TEST PROCEDURE
Appendix A is the form on which to report the results of an evaluation of a pipeline leak
detection system conducted according to the EPA test procedure. There are three
variants of this form. The choice depends on whether the leak detection system is used
as a line tightness test, a monthly monitoring test, or an hourly test. Use the variant that
is appropriate for the system you have evaluated. If the system was evaluated as all
three or any combination of these, fill out each variant that is applicable.
The appropriate variant of this form is to be filled out by the evaluating organization upon
completion of the evaluation of the system. All items are to be filled out and the
appropriate boxes checked. If a question is not applicable to the system, write "NA" in
the appropriate space. In addition, there are seven attachments that must be filled out.
104
-------
Results of the Performance Evaluation
Conducted According to EPA Test Procedures
Pipeline Leak Detection System
Used as a
Line Tightness Test
This form summarizes the results of an evaluation to determine whether the pipeline leak detection
system named below and described in Attachment 1 complies with federal regulations for conducting a
line tightness test. The evaluation was conducted according to the United States Environmental
Protection Agency's (EPA's) evaluation procedure, specified in Standard Test Procedures for
Evaluating Leak Detection Methods: Pipeline Leak Detection Systems. The full evaluation report
includes seven attachments.
Tank system owners who use this pipeline leak detection system should keep this form on file to show
compliance with the federal regulations. Tank system owners should check with state and local
agencies to make sure this form satisfies the requirements of these agencies.
System Evaluated
System Name:
Version of System:
Manufacturer Name:
(street address)
(city, state, zip code)
(telephone number)
Evaluation Results
1. The performance of this system
( ) meets or exceeds
( ) does not meet
the federal standards established by the EPA regulation for line tightness tests.
The EPA regulation for a line tightness test requires that the system be capable of detecting a
leak as small as 0.1 gal/h with a probability of detection (PD) of 95% and a probability of false
alarm (PFA) of 5%.
2. The estimated PFA in this evaluation is %and the estimated PD against a leak rate of 0.1
gal/h defined at a pipeline pressure of 20 psi in this evaluation is %.
Criterion for Declaring a Leak
Pipeline Leak Detection System - Results Form Page 1 of 5
-------
3. This system
( ) uses a preset threshold
( ) measures and reports the output quantity and compares it to a predetermined threshold to
determine whether the pipeline is leaking.
4. This system
( ) uses a single test
( ) uses a multiple-test sequence consisting of tests (specify number of tests
required) separated by hours (specify the time interval between tests) to
determine whether the pipeline is leaking.
5. This system declares a leak if the output of the measurement system exceeds a threshold of
(specify flow rate in gal/h) in out of tests (specify, for example, 1 out of
2, 2 out of 3). Please give additional details, if necessary, in the space provided.
Evaluation Approach
6. There are five options for collecting the data used in evaluating the performance of this system.
This system was evaluated
( ) at a special test facility (Option 1)
( ) at one or more instrumented operational storage tank facilities (Option 2)
( ) at five or more operational storage tank facilities verified to be tight (Option 3)
( ) at 10 or more operational storage tank facilities (Option 4)
( ) with an experimentally validated computer simulation (Option 5)
7. A total of tests were conducted on nonleaking tank(s) between (date)
and (date). A description of the pipeline configuration used in the evaluation is given
in Attachment 3.
/Answer questions 8 and 9 if Option 1, 2, or 5 was used.
8. The pipeline used in the evaluation was in. in diameter, ft long and
constructed of (fiberglass, steel, or other).
9. A mechanical line leak detector
( ) was
( ) was not
present in the pipeline system.
/Answer questions 10 and 11 if Option 3 or 4 was used.
10. The evaluation was conducted on (how many) pipeline systems ranging in diameter
from in. to in., ranging in length from ft to ft, and
constructed of (specify materials).
Pipeline Leak Detection System - Results Form Page 2 of 5
-------
11. A mechanical line leak detector
( ) was
( ) was not
present in the majority of the pipeline systems used in the evaluation.
12. Please specify how much time elapsed between the delivery of product and the start of the data
collection:
( ) 0 to 6 h
( ) 6 to 12 h
( ) 12 to 24 h
( ) 24 h or more
Temperature Conditions
This system was evaluated under the range of temperature conditions specified in Table 1. The
difference between the temperature of the product circulated through the pipeline for 1 h or more and
the average temperature of the backfill and soil between 2 and 12 in. from the pipeline is summarized in
Table 1. If Option 1, 2 or 5 was used a more detailed summary of the product temperature conditions
generated for the evaluation is presented in Attachment 4. If Option 3 or 4 was used, no artificial
temperature conditions were generated.
Table 1. Summary of Temperature Conditions Used in the Evaluation
Minimum Number
of Conditions Required
1
4
5
5
5
4
1
Number of Conditions Used*
Range of AT (°F)**
AT < -25
-25 < AT < -15
-15 25
* This column should be filled out only if Option 1, 2, or 5 was used.
** AT is the difference between the temperature of the product dispensed through the pipeline for over an hour
prior to the conduct of a test and the average temperature of the backfill and soil surrounding the pipe.
Data Used to Make Performance Estimates
13. The induced leak rate and the test results used to estimate the performance of this system are
summarized in Attachment 5. Were any test runs removed from the data set?
( ) no
( ) yes
If yes, please specify the reason and include with Attachment 5. (If more than one test was
removed, specify each reason separately.)
Pipeline Leak Detection System - Results Form
Page 3 of 5
-------
Sensitivity to Trapped Vapor
14. ( ) According to the vendor, this system can be used even if trapped vapor is present in the
pipeline during a test.
( ) According to the vendor, this system should not be used if trapped vapor is present in the
pipeline.
15. The sensitivity of this system to trapped vapor is indicated by the test results summarized in
Table 2. These tests were conducted at psi with ml of vapor trapped in the
line at a pressure of 0 psi. The data and test conditions are reported in Attachment 6.
Table 2. Summary of the Results of Trapped Vapor Tests
Test No.
1
2
3
AT
Induced Leak Rate
(gal/h)
Measured Leak Rate
(gal/h)
Performance Characteristics of the Instrumentation
16. State below the performance characteristics of the primary measurement system(s) used to
collect the data. (Please specify the units, for example, gallons, inches.)
Quantity Measured:
Resolution:
Precision
Accuracy:
Minimum Detectable Quantity:
Response Time:
Threshold is exceeded when the flow rate due to a leak exceeds
, gal/h.
Application of the System
17. This leak detection system is intended to test pipeline systems that are associated with
underground storage tank facilities, that contain petroleum or other chemical products, that are
typically constructed of fiberglass or steel, and that typically measure 2 in. in diameter and 200 ft
or less in length. The performance estimates are valid when:
• the system that was evaluated has not been substantially changed by subsequent
modifications
• the manufacturer's instructions for using the system are followed
• a mechanical line leak detector
( ) is present in
( ) has been removed from
the pipeline (check both if appropriate)
Pipeline Leak Detection System - Results Form
Page 4 of 5
-------
the waiting time between the last delivery of product to the underground storage tank
and the start of data collection for the test is h
the waiting time between the last dispensing of product through the pipeline system and
the start of data collection for the test is h
the total data collection time for the test is h
• the volume of the product in the pipeline system is less than twice the volume of the
product in the pipeline system used in the evaluation, unless a separate written
justification for testing larger pipeline systems is presented by the manufacturer,
concurred with by the evaluator, and attached to this evaluation as Attachment 8
• please give any other limitations specified by the vendor or determined during the
evaluation:
Disclaimer: This test procedure only addresses the issue of the system's ability to detect leaks in
pipelines. It does not rest the equipment for safety hazards or assess the operational functionality,
reliability or maintainability of the equipment.
Attachments
Attachment 1 - Description of the System Evaluated
Attachment 2 - Summary of the Performance of the System Evaluated
Attachment 3 - Summary of the Configuration of the Pipeline System(s) Used in the Evaluation
Attachment 4 - Data Sheet Summarizing Product Temperature Conditions Used in the Evaluation
Attachment 5 - Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Attachment 6 - Data Sheet Summarizing the Test Results and the Trapped Vapor Tests
Attachment 7 - Data Sheet Summarizing the Test Results Used to Check the Relationship Supplied by
the Manufacturer for Combining the Signal and Noise
Certification of Results
I certify that the pipeline leak detection system was operated according to the vendor's instructions. I
also certify that the evaluation was performed according to the procedure specified by the EPA and that
the results presented above are those obtained during the evaluation.
(name of person performing evaluation) (organization performing evaluation)
(signature) (street address)
(date) (city, state, zip)
(telephone number)
Pipeline Leak Detection System - Results Form Page 5 of 5
-------
Results of the Performance Evaluation
Conducted According to EPA Test Procedures
Pipeline Leak Detection System
Used as a
Monthly Monitoring Test
This form summarizes the results of an evaluation to determine whether the pipeline leak detection
system named below and described in Attachment 1 complies with federal regulations for conducting a
monthly monitoring test. The evaluation was conducted according to the United States Environmental
Protection Agency's (EPA's) evaluation procedure, specified in Standard Test Procedures for
Evaluating Leak Detection Methods: Pipeline Leak Detection Systems. The full evaluation report
includes seven attachments.
Tank system owners who use this pipeline leak detection system should keep this form on file to show
compliance with the federal regulations. Tank system owners should check with state and local
agencies to make sure this form satisfies the requirements of these agencies.
System Evaluated
System Name:
Version of System:
Manufacturer Name:
(street address)
(city, state, zip code)
(telephone number)
Evaluation Results
1. The performance of this system
( ) meets or exceeds
( ) does not meet
the federal standards established by the EPA regulation for monthly monitoring tests.
The EPA regulation for a monthly monitoring test requires that the system be capable of
detecting a leak as small as 0.2 gal/h with a probability of detection (PD) of 95% and a
probability of false alarm (PFA) of 5%.
2. The estimated PFA in this evaluation is % and the estimated PD against a leak rate
of 0.2 gal/h defined at a pipeline pressure of 20 psi in this evaluation is %.
Pipeline Leak Detection System - Results Form Page 1 of 5
-------
Criterion for Declaring a Leak
3. This system
( ) uses a preset threshold
( ) measures and reports the output quantity and compares it to a predetermined threshold to
determine whether the pipeline is leaking.
4. This system
( ) uses a single test
( ) uses a multiple-test sequence consisting of tests (specify number of tests
required) separated by hours (specify the time interval between tests) to
determine whether the pipeline is leaking.
5. This system declares a leak if the output of the measurement system exceeds a threshold of
(specify flow rate in gal/h) in out of tests (specify, for
example, 1 out of 2, 2 out of 3). Please give additional details, if necessary, in the space
provided.
Evaluation Approach
6. There are five options for collecting the data used in evaluating the performance of this system.
This system was evaluated
( ) at a special test facility (Option 1)
( ) at one or more instrumented operational storage tank facilities (Option 2)
( ) at five or more operational storage tank facilities verified to be tight (Option 3)
( ) at 10 or more operational storage tank facilities (Option 4)
( ) with an experimentally validated computer simulation (Option 5)
7. A total of tests were conducted on nonleaking tank(s) between (date)
and (date). A description of the pipeline configuration used in the
evaluation is .given in Attachment 3.
/Answer questions 8 and 9 if Option 1, 2, or 5 was used.
8. The pipeline used in the evaluation was in. in diameter, ft long and
constructed of (fiberglass, steel, or other).
9. A mechanical line leak detector
( ) was
( ) was not
present in the pipeline system.
/Answer questions 10 and 11 if Option 3 or 4 was used.
10. The evaluation was conducted on (how many) pipeline systems ranging in diameter
from in. to in., ranging in length from ft to ft, and
constructed of (specify materials).
Pipeline Leak Detection System - Results Form Page 2 of 5
-------
11. A mechanical line leak detector
( ) was
( ) was not
present in the majority of the pipeline systems used in the evaluation.
12. Please specify how much time elapsed between the delivery of product and the start of the data
collection:
( ) 0 to 6 h
( ) 6 to 12 h
( ) 12 to 24 h
( ) 24 h or more
Temperature Conditions
This system was evaluated under the range of temperature conditions specified in Table 1. The
difference between the product circulated through the pipeline for 1 h or more and the average
temperature of the backfill and soil between 2 and 12 in. from the pipeline is summarized in Table 1. If
Option 1, 2 or 5 was used a more detailed summary of the product temperature conditions generated
for the evaluation is presented in Attachment 4. If Option 3 or 4 was used, no artificial temperature
conditions were generated.
Table 1. Summary of Temperature Conditions Used in the Evaluation
Minimum Number
of Conditions Required
1
4
5
5
5
4
1
Number of Conditions Used*
Range of AT (°F)**
AT < -25
-25 < AT < -15
-1525
* This column should be filled out only if Option 1, 2, or 5 was used.
** AT is the difference between the temperature of the product dispensed through the pipeline for over an hour
prior to the conduct of a test and the average temperature of the backfill and soil surrounding the pipe.
Data Used to Make Performance Estimates
13. The induced leak rate and the test results used to estimate the performance of this system are
summarized in Attachment 5. Were any test runs removed from the data set?
( ) no
( ) yes
If yes, please specify the reason and include with Attachment 5. (If more than one test was
removed, specify each reason separately.)
Pipeline Leak Detection System - Results Form
Page 3 of 5
-------
Sensitivity to Trapped Vapor
14. ( ) According to the vendor, this system can be used even if trapped vapor is present in the
pipeline during a test.
( ) According to the vendor, this system should not be used if trapped vapor is present in the
pipeline.
15. The sensitivity of this system to trapped vapor is indicated by the test results summarized in
Table 2. These tests were conducted at psi with ml of vapor trapped
in the line at a pressure of 0 psi. The data and test conditions are reported in Attachment 6.
Table 2. Summary of the Results of Trapped Vapor Tests
Test No.
1
2
3
AT
(°F)
Induced Leak Rate
(gal/h)
Measured Leak Rate
(gal/h)
Performance Characteristics of the Instrumentation
16. State below the performance characteristics of the primary measurement system used to collect
the data. (Please specify the units, for example, gallons, inches.)
Quantity Measured:
Resolution:
Precision
Accuracy:
Minimum Detectable Quantity:
Response Time:
Threshold is exceeded when the flow rate due to a leak exceeds
. gal/h.
Application of the System
17. This leak detection system is intended to test pipeline systems that are associated with
underground storage tank facilities, that contain petroleum or other chemical products, that are
typically constructed of fiberglass or steel, and that typically measure 2 in. in diameter and 200 ft
or less in length. The performance estimates are valid when:
• the system that was evaluated has not been substantially changed by subsequent
modifications
• the manufacturer's instructions for using the system are followed
• a mechanical line leak detector
( ) is present in
( ) has been removed from
the pipeline (check both if appropriate)
Pipeline Leak Detection System - Results Form
Page 4 of 5
-------
the waiting time between the last delivery of product to the underground storage tank
and the start of data collection for the test is h
the waiting time between the last dispensing of product through the pipeline system and
the start of data collection for the test is h
the total data collection time for the test is h
• the volume of the product in the pipeline system is less than twice the volume of the
product in the pipeline system used in the evaluation, unless a separate written
justification for testing larger pipeline systems is presented by the manufacturer,
concurred with by the evaluator, and attached to this evaluation as Attachment 8
• please give any other limitations specified by the vendor or determined during the
evaluation:
Disclaimer: This test procedure only addresses the issue of the system's ability to detect leaks in
pipelines. It does not rest the equipment for safety hazards or assess the operational functionality,
reliability or maintainability of the equipment.
Attachments
Attachment 1 - Description of the System Evaluated
Attachment 2 - Summary of the Performance of the System Evaluated
Attachment 3 - Summary of the Configuration of the Pipeline System(s) Used in the Evaluation
Attachment 4 - Data Sheet Summarizing Product Temperature Conditions Used in the Evaluation
Attachment 5 - Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Attachment 6 - Data Sheet Summarizing the Test Results and the Trapped Vapor Tests
Attachment 7 - Data Sheet Summarizing the Test Results Used to Check the Relationship Supplied by
the Manufacturer for Combining the Signal and Noise
Certification of Results
I certify that the pipeline leak detection system was operated according to the vendor's instructions. I
also certify that the evaluation was performed according to the procedure specified by the EPA and that
the results presented above are those obtained during the evaluation.
(name of person performing evaluation) (organization performing evaluation)
(signature) (street address)
(date) (city, state, zip)
(telephone number)
Pipeline Leak Detection System - Results Form Page 5 of 5
-------
Results of the Performance Evaluation
Conducted According to EPA Test Procedures
Pipeline Leak Detection System
Used as a
Hourly Test
This form summarizes the results of an evaluation to determine whether the pipeline leak detection
system named below and described in Attachment 1 complies with federal regulations for conducting
an hourly test. The evaluation was conducted according to the United States Environmental Protection
Agency's (EPA's) evaluation procedure, specified in Standard Test Procedures for Evaluating Leak
Detection Methods: Pipeline Leak Detection Systems. The full evaluation report includes seven
attachments.
Tank system owners who use this pipeline leak detection system should keep this form on file to show
compliance with the federal regulations. Tank system owners should check with state and local
agencies to make sure this form satisfies the requirements of these agencies.
System Evaluated
System Name:
Version of System:
Manufacturer Name:
(street address)
(city, state, zip code)
(telephone number)
Evaluation Results
1. The performance of this system
( ) meets or exceeds
( ) does not meet
the federal standards established by the EPA regulation for hourly tests.
The EPA regulation for an hourly test requires that the system be capable of detecting a leak as
small as 3.0 gal/h with a probability of detection (PD) of 95% and a probability of false alarm
(PFA) of 5%.
2. The estimated PFA in this evaluation is % and the estimated PD against a leak rate
of 3.0 gal/h defined at a pipeline pressure of 20 psi in this evaluation is %.
Pipeline Leak Detection System - Results Form Page 1 of 5
-------
Criterion for Declaring a Leak
3. This system
( ) uses a preset threshold
( ) measures and reports the output quantity and compares it to a predetermined threshold to
determine whether the pipeline is leaking.
4. This system
( ) uses a single test
( ) uses a multiple-test sequence consisting of tests (specify number of tests
required) separated by hours (specify the time interval between tests) to
determine whether the pipeline is leaking.
5. This system declares a leak if the output of the measurement system exceeds a threshold of
(specify flow rate in gal/h) in out of tests (specify, for
example, 1 out of 2, 2 out of 3). Please give additional details, if necessary, in the space
provided.
Evaluation Approach
6. There are five options for collecting the data used in evaluating the performance of this system.
This system was evaluated
( ) at a special test facility (Option 1)
( ) at one or more instrumented operational storage tank facilities (Option 2)
( ) at five or more operational storage tank facilities verified to be tight (Option 3)
( ) at 10 or more operational storage tank facilities (Option 4)
( ) with an experimentally validated computer simulation (Option 5)
7. A total of tests were conducted on nonleaking tank(s) between (date)
and (date). A description of the pipeline configuration used in the evaluation is given
in Attachment 3.
/Answer questions 8 and 9 if Option 1, 2, or 5 was used.
8. The pipeline used in the evaluation was in. in diameter, ft long and
constructed of (fiberglass, steel, or other).
9. A mechanical line leak detector
( ) was
( ) was not
present in the pipeline system.
/Answer questions 10 and 11 if Option 3 or 4 was used.
10. The evaluation was conducted on (how many) pipeline systems ranging in diameter
from in. to in., ranging in length from ft to ft, and
constructed of (specify materials).
Pipeline Leak Detection System - Results Form Page 2 of 5
-------
11. A mechanical line leak detector
( ) was
( ) was not
present in the majority of the pipeline systems used in the evaluation.
12. Please specify how much time elapsed between the delivery of product and the start of the data
collection:
( ) 0 to 6 h
( ) 6 to 12 h
( ) 12 to 24 h
( ) 24 h or more
Temperature Conditions
This system was evaluated under the range of temperature conditions specified in Table 1. The
difference between the product circulated through the pipeline for 1 h or more and the average
temperature of the backfill and soil between 2 and 12 in. from the pipeline is summarized in Table 1. If
Option 1, 2 or 5 was used a more detailed summary of the product temperature conditions generated
for the evaluation is presented in Attachment 4. If Option 3 or 4 was used, no artificial temperature
conditions were generated.
Table 1. Summary of Temperature Conditions Used in the Evaluation
Minimum Number
of Conditions Required
1
4
5
5
5
4
1
Number of Conditions Used*
Range of AT (°F)**
AT < -25
-25 < AT < -15
-1525
* This column should be filled out only if Option 1, 2, or 5 was used.
** AT is the difference between the temperature of the product dispensed through the pipeline for over an hour
prior to the conduct of a test and the average temperature of the backfill and soil surrounding the pipe.
Data Used to Make Performance Estimates
13. The induced leak rate and the test results used to estimate the performance of this system are
summarized in Attachment 5. Were any test runs removed from the data set?
( ) no
( ) yes
If yes, please specify the reason and include with Attachment 5. (If more than one test was
removed, specify each reason separately.)
Pipeline Leak Detection System - Results Form
Page 3 of 5
-------
Sensitivity to Trapped Vapor
14. ( ) According to the vendor, this system can be used even if trapped vapor is present in the
pipeline during a test.
( ) According to the vendor, this system should not be used if trapped vapor is present in the
pipeline.
15. The sensitivity of this system to trapped vapor is indicated by the test results summarized in
Table 2. These tests were conducted at psi with ml of vapor trapped
in the line at a pressure of 0 psi. The data and test conditions are reported in Attachment 6.
Table 2. Summary of the Results of Trapped Vapor Tests
Test No.
1
2
3
AT
(°F)
Induced Leak Rate
(gal/h)
Measured Leak Rate
(gal/h)
Performance Characteristics of the Instrumentation
16. State below the performance characteristics of the primary measurement system(s) used to
collect the data. (Please specify the units, for example, gallons, inches.)
Quantity Measured:
Resolution:
Precision
Accuracy:
Minimum Detectable Quantity:
Response Time:
Threshold is exceeded when the flow rate due to a leak exceeds
, gal/h.
Application of the System
17. This leak detection system is intended to test pipeline systems that are associated with
underground storage tank facilities, that contain petroleum or other chemical products, that are
typically constructed of fiberglass or steel, and that typically measure 2 in. in diameter and 150 ft
or less in length. The performance estimates are valid when:
• the system that was evaluated has not been substantially changed by subsequent
modifications
• the manufacturer's instructions for using the system are followed
• a mechanical line leak detector
( ) is present in
( ) has been removed from
the pipeline(check both if appropriate)
Pipeline Leak Detection System - Results Form
Page 4 of 5
-------
the waiting time between the last delivery of product to the underground storage tank
and the start of data collection for the test is h
the waiting time between the last dispensing of product through the pipeline system and
the start of data collection for the test is h
the total data collection time for the test is h
• the volume of the product in the pipeline system is less than twice the volume of the
product in the pipeline system used in the evaluation, unless a separate written
justification for testing larger pipeline systems is presented by the manufacturer,
concurred with by the evaluator, and attached to this evaluation as Attachment 8
• please give any other limitations specified by the vendor or determined during the
evaluation:
Disclaimer: This test procedure only addresses the issue of the system's ability to detect leaks in
pipelines. It does not rest the equipment for safety hazards or assess the operational functionality,
reliability or maintainability of the equipment.
Attachments
Attachment 1 - Description of the System Evaluated
Attachment 2 - Summary of the Performance of the System Evaluated
Attachment 3 - Summary of the Configuration of the Pipeline System(s) Used in the Evaluation
Attachment 4 - Data Sheet Summarizing Product Temperature Conditions Used in the Evaluation
Attachment 5 - Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Attachment 6 - Data Sheet Summarizing the Test Results and the Trapped Vapor Tests
Attachment 7 - Data Sheet Summarizing the Test Results Used to Check the Relationship Supplied by
the Manufacturer for Combining the Signal and Noise
Certification of Results
I certify that the pipeline leak detection system was operated according to the vendor's instructions. I
also certify that the evaluation was performed according to the procedure specified by the EPA and that
the results presented above are those obtained during the evaluation.
(name of person performing evaluation) (organization performing evaluation)
(signature) (street address)
(date) (city, state, zip)
(telephone number)
Pipeline Leak Detection System - Results Form Page 5 of 5
-------
APPENDIX B
ATTACHMENTS 1 THROUGH 7 TO THE FORM IN APPENDIX A
Appendix B comprises the seven attachments to the form in Appendix A.
In Attachment 1 you are asked to describe the pipeline leak detection system by
answering 26 questions, most of which are multiple-choice.
In Attachment 2, you are asked to present a summary of performance estimates by filling
in the tables provided. Like the form in Appendix A, Attachment 2 has three variants,
depending on whether the leak detection system is used as a line tightness test, a
monthly monitoring test, or an hourly test. Choose the variant that is appropriate for the
system you have evaluated. In addition, if your system uses a multiple-test strategy,
please fill out that part of Attachment 2 which asks for the results of the first test in the
sequence.
In Attachment 3, you are asked to summarize the configuration of the pipeline systems
used in the evaluation. The charts that are provided are broken down according to the
options selected for the evaluation. For example, if the system was evaluated at a
specialized test facility, at an instrumented operational LIST facility, or by computer
simulation, fill out the chart marked "Options 1, 2 and 5." If the system was evaluated at
five operational LIST facilities whose integrity had been verified, fill in the chart marked
"Option 3." If the system was evaluated at 10 or more operational LIST facilities, use the
chart marked "Option 4."
In Attachment 4, you are asked to summarize the temperature conditions used in the
evaluation. Again, the charts are broken down according to the options selected for the
evaluation.
In Attachments 5 and 6, you are asked to summarize the leak rates and the trapped
vapor tests, respectively. You are also asked to summarize the results of the tests
performed. The charts provided are organized similarly to those in Attachment 4.
In Attachment 7, you are asked to summarize the test results that are used to check the
relationship provided by the manufacturer, which describes how the signal adds to the
noise.
120
-------
Attachment 1
Description
Pipeline Leak Detection System
This form provides supporting information on the operating principles of the leak detection system or on
how the equipment works. This form is to be filled out by the evaluating organization with assistance
from the manufacturer before the start of the evaluation.
Describe the important features of the system as indicated below. A detailed description is not
required, nor is it necessary to reveal proprietary features of the system.
To minimize the time required to complete this form, the most frequently expected answers to the
questions have been provided. For those answers that are dependent on site conditions, please give
answers that apply in "typical" conditions. Please write in any additional information about the system
that you believe is important.
Check all appropriate boxes for each question. Check more than one box per question if it applies. If
'Other' is checked, please complete the space provided to specify or briefly describe the matter. If
necessary, use all the white space next to a question to complete a description.
System Name and Version:
Date:
Applicability of the System
1. With what products can this system be used? (Check all applicable responses.)
( ) gasoline
( ) diesel
( ) aviation fuel
( ) fuel oil #4
( ) fuel oil #6
( ) solvent
( ) waste oil
( ) other (specify)
2. What types of pipelines can be tested? (Check all applicable responses.)
( ) fiberglass
( ) steel
( ) other (specify)
3. Can this leak detection system be used to test double-wall pipeline systems?
( ) yes ( ) no
Description - Pipeline Leak Detection System Page 1 of 5
-------
4. What is the nominal diameter of a pipeline that can be tested with this system?
( ) 1 in. or less
( ) between 1 and 3 in.
( ) between 3 and 6 in.
( ) between 6 and 10 in.
( ) other
5. The system can be used on pipelines pressurized to psi.
The safe maximum operating pressure for this system is psi.
6. Does the system conduct a test while a mechanical line leak detector is in place in the pipeline?
( ) yes ( ) no
General Features of the System
7. What type of test is the system conducting? (Check all applicable responses.)
( ) 0.1 gal/h Line Tightness Test
( ) 0.2 gal/h Monthly Monitoring Test
( ) 3 gal/h Hourly Test
8. Is the system permanently installed on the pipeline?
( ) yes ( ) no
Does the system test the line automatically?
( ) yes ( ) no
If a leak is declared, what does the system do? (Check all applicable responses.)
( ) displays or prints a message
( ) triggers an alarm
( ) alerts the operator
( ) shuts down the dispensing system
9. What quantity or quantities are measured by the system? (Please list.)
10. Does the system use a preset threshold that is automatically activated or that automatically
turns on an alarm?
( ) yes (If yes, skip question 11.)
( ) no (If no, answer question 11.)
11. Does the system measure and report the quantity
( ) yes ( ) no
Description - Pipeline Leak Detection System Page 2 of 5
-------
If so, is the output quantity converted to flow rate in gallons per hour?
( ) yes ( ) no
12. What is the specified line pressure during a test?
( ) operating pressure of line
( ) 150% of operating pressure
( ) a specific test pressure of psi
Test Protocol
13. What is the minimum waiting period required between a delivery of product to an underground
storage tank and the start of the data collection for a pipeline leak detection test?
( ) no waiting period
( ) less than 15 min
( ) 15 min to 1 h
( ) 1 to 5 h
( ) 6 to 12 h
( ) 12 to 24 h
( ) greater than 24 h
( ) variable (Briefly explain.)
14. What is the minimum waiting period required between the last dispensing of product through the
pipeline and the start of the data collection for a pipeline leak detection test?
( ) no waiting period
( ) less than 15 min
( ) 15 min to 1 h
( ) 1 to 4 h
( ) 4 to 8 h
( ) greater than 8 h
( ) variable (Briefly explain.)
15. What is the minimum amount of time necessary to set up equipment and complete a leak
detection test? (Include setup time, waiting time and data collection time. If a multiple-test
sequence is used, give the amount of time necessary to complete the first test as well as the
total amount of time necessary to complete the entire sequence.)
h (single test)
h (multiple test)
16. Does the system compensate for those pressure or volume changes of the product in the
pipeline that are due to temperature changes?
( ) yes ( ) no
17. Is there a special test to check the pipeline for trapped vapor?
( ) yes ( ) no
Description - Pipeline Leak Detection System Page 3 of 5
-------
18. Can a test be performed with trapped vapor in the pipeline?
( ) yes ( ) no
19. If trapped vapor is found in the pipeline, is it removed before a test is performed?
( ) yes ( ) no
20. Are deviations from this protocol acceptable?
( ) yes ( ) no
If yes, briefly specify:
21. Are elements of the test procedure determined by on-site testing personnel?
( ) yes ( ) no
If yes, which ones? (Check all applicable responses.)
( ) waiting period between filling the tank and the beginning of data collection for the test
( ) length of test
( ) determination of the presence of vapor pockets
( ) determination of "outlier" (or anomalous) data that may be discarded
( ) other (Describe briefly.)
Data Acquisition
22. How are the test data acquired and recorded?
( ) manually
( ) by strip chart
( ) by computer
( ) by microprocessor
23. Certain calculations are necessary to reduce and analyze the data. How are these calculations
done?
( ) manual calculations by the operator on site
( ) interactive computer program used by the operator
( ) automatically done with a computer program
( ) automatically done with a microprocessor
Detection Criterion
24. What threshold is used to determine whether the pipeline is leaking?
(in the units used by the measurement system)
(in gal/h)
Description - Pipeline Leak Detection System Page 4 of 5
-------
25. Is a multiple-test sequence used to determine whether the pipeline is leaking?
( ) yes (If yes, answer the three questions below)
( ) no (If no, skip the three questions below)
How many tests are conducted?
How many tests are required before a leak can be declared?
What is the time between tests?
(Enter 0 if the tests are conducted one after the other.)
Calibration
26. How frequently are the sensor systems calibrated?
( ) never
( ) before each test
( ) weekly
( ) monthly
( ) semi-annually
( ) yearly or less frequently
Description - Pipeline Leak Detection System Page 5 of 5
-------
Attachment 2
Summary of Performance Estimates
Pipeline Leak Detection System
Used as a
Line Tightness Test
Complete this page if the pipeline leak detection system has been evaluated as a line tightness test.
Please complete the first table. Completion of the last three tables is optional. (The last three tables
present the performance of the system for different combinations of thresholds, probabilities of false
alarm, and probabilities of detection. They are useful for comparing the performance of this system to
that of other systems.)
Performance of the Pipeline Leak Detection System as Evaluated
Description
Evaluated System
EPA Standard
Leak Rate
(gal/h)
0.10
0.10
PD
0.95
PFA
0.05
Threshold
(gal/h)
N/A
Probability of False Alarm as a Function of Threshold
Threshold
(gal/h)
Probability of False Alarm
0.10
0.075
0.05
0.05
Probability of Detection as a Function of Threshold for a Leak Rate of 0.10 gal/h
Threshold
(gal/h)
Probability of Detection
0.95
0.90
0.80
0.50
Smallest Leak Rate That Can Be Detected with the Specified Probability of Detection
and Probability of False Alarm
Leak Rate
(gal/h)
Probability of Detection
0.95
0.95
0.95
0.90
0.80
0.50
Probability of False Alarm
0.10
0.075
0.05
0.05
0.05
0.05
-------
Summary of Performance Estimates
Pipeline Leak Detection System
Used as a
Line Tightness Test
First Test of a Multiple-Test Sequence
Complete these tables only if the system being evaluated requires, as part of its test protocol, more
than one complete test to determine whether the pipeline is leaking. System performance based on the
first test alone must be reported on this form. Please complete the first table. Completion of the last
three tables is optional. (The last three tables present the performance of the system for different
combinations of thresholds, probabilities of false alarm, and probabilities of detection. They are useful
for comparing the performance of this system to that of other systems.)
Performance of the Pipeline Leak Detection System as Evaluated
Description
Evaluated System
EPA Standard
Leak Rate
(gal/h)
0.10
0.10
PD
0.95
PFA
0.05
Threshold
(gal/h)
N/A
Probability of False Alarm as a Function of Threshold
Threshold
(gal/h)
Probability of False Alarm
0.10
0.075
0.05
0.05
Probability of Detection as a Function of Threshold for a Leak Rate of 0.10 gal/h
Threshold
(gal/h)
Probability of Detection
0.95
0.90
0.80
0.50
Smallest Leak Rate That Can Be Detected with the Specified Probability of Detection
and Probability of False Alarm
Leak Rate
(gal/h)
Probability of Detection
0.95
0.95
0.95
0.90
0.80
0.50
Probability of False Alarm
0.10
0.075
0.05
0.05
0.05
0.05
-------
Attachment 2
Summary of Performance Estimates
Pipeline Leak Detection System
Used as a
Monthly Monitoring Test
Complete this page if the pipeline leak detection system has been evaluated as a monthly monitoring
test. Please complete the first table. Completion of the last three tables is optional. (The last three
tables present the performance of the system for different combinations of thresholds, probabilities of
false alarm, and probabilities of detection. They are useful for comparing the performance of this
system to that of other systems.)
Performance of the Pipeline Leak Detection System as Evaluated
Description
Evaluated System
EPA Standard
Leak Rate
(gal/h)
0.20
0.20
PD
0.95
PFA
0.05
Threshold
(gal/h)
N/A
Probability of False Alarm as a Function of Threshold
Threshold
(gal/h)
Probability of False Alarm
0.10
0.075
0.05
0.05
Probability of Detection as a Function of Threshold for a Leak Rate of 0.20 gal/h
Threshold
(gal/h)
Probability of Detection
0.95
0.90
0.80
0.50
Smallest Leak Rate that Can Be Detected with the Specified Probability of Detection
and Probability of False Alarm
Leak Rate
(gal/h)
Probability of Detection
0.95
0.95
0.95
0.90
0.80
0.50
Probability of False Alarm
0.10
0.075
0.05
0.05
0.05
0.05
-------
Summary of Performance Estimates
Pipeline Leak Detection System
Used as a
Monthly Monitoring Test
First Test of a Multiple-Test Sequence
Complete these tables only if the system being evaluated requires, as part of its test protocol, more
than one complete test to determine whether the pipeline is leaking. System performance based on the
first test alone must be reported on this form. Please complete the first table. Completion of the last
three tables is optional. (The last three tables present the performance of the system for different
combinations of thresholds, probabilities of false alarm, and probabilities of detection. They are useful
for comparing the performance of this system to that of other systems.)
Performance of the Pipeline Leak Detection System as Evaluated
Description
Evaluated System
EPA Standard
Leak Rate
(gal/h)
0.20
0.20
PD
0.95
PFA
0.05
Threshold
(gal/h)
N/A
Probability of False Alarm as a Function of Threshold
Threshold
(gal/h)
Probability of False Alarm
0.10
0.075
0.05
0.05
Probability of Detection as a Function of Threshold for a Leak Rate of 0.20 gal/h
Threshold
(gal/h)
Probability of Detection
0.95
0.90
0.80
0.50
Smallest Leak Rate That Can Be Detected with the Specified Probability of Detection
and Probability of False Alarm
Leak Rate
(gal/h)
Probability of Detection
0.95
0.95
0.95
0.90
0.80
0.50
Probability of False Alarm
0.10
0.075
0.05
0.05
0.05
0.05
-------
Attachment 2
Summary of Performance Estimates
Pipeline Leak Detection System
Used as an
Hourly Test
Complete this page if the pipeline leak detection system has been evaluated as an hourly test. Please
complete the first table. Completion of the last three tables is optional. (The last three tables present
the performance of the system for different combinations of thresholds, probabilities of false alarm, and
probabilities of detection. They are useful for comparing the performance of this system to that of other
systems.)
Performance of the Pipeline Leak Detection System as Evaluated
Description
Evaluated System
EPA Standard
Leak Rate
(gal/h)
3.0
3.0
PD
0.95
PFA
0.05
Threshold
(gal/h)
N/A
Probability of False Alarm as a Function of Threshold
Threshold
(gal/h)
Probability of False Alarm
0.10
0.075
0.05
0.05
Probability of Detection as a Function of Threshold for a Leak Rate of 3.0 gal/h
Threshold
(gal/h)
Probability of Detection
0.95
0.90
0.80
0.50
Smallest Leak Rate That Can Be Detected with the Specified Probability of Detection
and Probability of False Alarm
Leak Rate
(gal/h)
Probability of Detection
0.95
0.95
0.95
0.90
0.80
0.50
Probability of False Alarm
0.10
0.075
0.05
0.05
0.05
0.05
-------
Summary of Performance Estimates
Pipeline Leak Detection System
Used as an
Hourly Test
First Test of a Multiple-Test Sequence
Complete this page only if the system being evaluated requires, as part of its test protocol, more than
one complete test to determine whether the pipeline is leaking. System performance based on the first
test alone must be reported on this form. Please complete the first table. Completion of the last three
tables is optional. (The last three tables present the performance of the system for different
combinations of thresholds, probabilities of false alarm, and probabilities of detection. They are useful
for comparing the performance of this system to that of other systems.)
Performance of the Pipeline Leak Detection System as Evaluated
Description
Evaluated System
EPA Standard
Leak Rate
(gal/h)
3.0
3.0
PD
0.95
PFA
0.05
Threshold
(gal/h)
N/A
Probability of False Alarm as a Function of Threshold
Threshold
(gal/h)
Probability of False Alarm
0.10
0.075
0.05
0.05
Probability of Detection as a Function of Threshold for a Leak Rate of 3.0 gal/h
Threshold
(gal/h)
Probability of Detection
0.95
0.90
0.80
0.50
Smallest Leak Rate That Can Be Detected with the Specified Probability of Detection
and Probability of False Alarm
Leak Rate
(gal/h)
Probability of Detection
0.95
0.95
0.95
0.90
0.80
0.50
Probability of False Alarm
0.10
0.075
0.05
0.05
0.05
0.05
-------
Attachment 3
Summary of the Configuration of the Pipeline System(s)
Used in the Evaluation
Pipeline Leak Detection System
Options 1, 2, and 5
Specialized Test Facility, Operational Storage Tank System, or Computer Simulation
Inside diameter of pipeline (in.)
Length of pipeline (tank to dispenser) (ft)
Volume of product in line during testing (gal)
Type of material (fiberglass, steel, other1)
Type of product in tank and pipeline (gasoline, diesel, other2)
Was a mechanical line leak detector present? (yes or no)
Was trapped vapor present? (yes or no)
Bulk Modulus (B) (psi)
BA/o (psi/ml)
Storage tank capacity (gal)
1 Specify type of construction material.
2 Specify type of product for each tank.
-------
Attachment 3
Summary of the Configuration of the Pipeline System(s)
Used in the Evaluation
Pipeline Leak Detection System
Option 3
Operational Tank System
Inside diameter of pipeline (in.)
Length of pipeline (tank to dispenser) (ft)
Volume of product in line during testing (gal)
Type of material (fiberglass, steel, other1)
Type of product in tank and pipeline (gasoline, diesel,
other2)
Was a mechanical line leak detector present? (yes or
no)
Was trapped vapor present? (yes or no)
Bulk Modulus (B) (psi)
BA/o (psi/ml)
Storage tank capacity (gal)
1
2
3
4
5
1 Specify type of construction material.
2 Specify type of product for each tank.
Operational Tank System
Inside diameter of pipeline (in.)
Length of pipeline (tank to dispenser) (ft)
Volume of product in line during testing (gal)
Type of material (fiberglass, steel, other1)
Type of product in tank and pipeline (gasoline, diesel,
other2)
Was a mechanical line leak detector present? (yes or
no)
Was trapped vapor present? (yes or no)
Bulk Modulus (B) (psi)
BA/o (psi/ml)
Storage tank capacity (gal)
6
7
8
9
10
1 Specify type of construction material.
2 Specify type of product for each tank.
-------
Attachment 3
Summary of the Configuration of the Pipeline System(s)
Used in the Evaluation
Pipeline Leak Detection System
Option 4
Operational Tank System
Inside diameter of pipeline (in.)
Length of pipeline (tank to dispenser) (ft)
Volume of product in line during testing (gal)
Type of material (fiberglass, steel, other1)
Type of product in tank and pipeline (gasoline, diesel,
other2)
Was a mechanical line leak detector present? (yes or
no)
Was trapped vapor present? (yes or no)
Bulk Modulus (B) (psi)
BA/o (psi/ml)
Storage tank capacity (gal)
1
2
3
4
5
Specify type of construction material.
! Specify type of product for each tank.
Operational Tank System
Inside diameter of pipeline (in.)
Length of pipeline (tank to dispenser) (ft)
Volume of product in line during testing (gal)
Type of material (fiberglass, steel, other1)
Type of product in tank and pipeline (gasoline, diesel,
other2)
Was a mechanical line leak detector present? (yes or
no)
Was trapped vapor present? lye's or no)
Bulk Modulus (B) (psi)
BA/o (psi/ml)
Storage tank capacity (gal)
6
7
8
9
10
Specify type of construction material.
! Specify type of product for each tank.
-------
Attachment 3
(Concluded)
Summary of the Configuration of the Pipeline System(s)
Used in the Evaluation
Pipeline Leak Detection System
Option 4
Operational Tank System
Inside diameter of pipeline (in.)
Length of pipeline (tank to dispenser) (ft)
Volume of product in line during testing (gal)
Type of material (fiberglass, steel, other1)
Type of product in tank and pipeline (gasoline, diesel,
other2)
Was a mechanical line leak detector present? (yes or
no)
Was trapped vapor present? (yes or no)
Bulk Modulus (B) (psi)
BA/o (psi/ml)
Storage tank capacity (gal)
11
12
13
14
15
1 Specify type of construction material.
2 Specify type of product for each tank.
Operational Tank System
Inside diameter of pipeline (in.)
Length of pipeline (tank to dispenser) (ft)
Volume of product in line during testing (gal)
Type of material (fiberglass, steel, other1)
Type of product in tank and pipeline (gasoline, diesel,
other2)
Was a mechanical line leak detector present? (yes or
no)
Was trapped vapor present? lye's or no)
Bulk Modulus (B) (psi)
BA/o (psi/ml)
Storage tank capacity (gal)
16
17
18
19
20
1 Specify type of construction material.
2 Specify type of product for each tank.
-------
Attachment 4
Data Sheet Summarizing the Product Temperature Conditions Used in the Evaluation
Pipeline Leak Detection System
Options 1 and 5
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Date
Test
Began
(D-M-Y)
Nominal
Product
Temperature
before
Circulation
Was Started
(°F)
Time
Circulation
Started
(local
military)
Time
Circulation
Ended
(°F)
Duration of
Circulation
(h-min)
Time of
Temperature
Measurements
(local military)
TTB
(°F)
Ti
(°F)
T2
(°F)
T3
(°F)
TG
(°F)
TTB -To
(°F)
Temperature
Test
Matrix
Category
(Table 5.1)
-------
Attachment 4
(concluded)
Data Sheet Summarizing the Product Temperature Conditions Used in the Evaluation
Pipeline Leak Detection System
Options 1 and 5
Test No.
(Based on
Temperature
Condition)
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Date
Test
Began
(D-M-Y)
Nominal
Product
Temperature
before
Circulation
Was Started
(°F)
Time
Circulation
Started
(local
military)
Time
Circulation
Ended
(°F)
Duration of
Circulation
(h-min)
Time of
Temperature
Measurements
(local military)
TTB
(°F)
Ti
(°F)
T2
(°F)
T3
(°F)
TG
(°F)
TTB -To
(°F)
Temperature
Test
Matrix
Category
(Table 5.1)
-------
Attachment 4
Data Sheet Summarizing the Product Temperature Conditions Used in the Evaluation
Pipeline Leak Detection System
Option 2
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Date
Test
Began
(D-M-Y)
Date of
Last
Product
Delivery
(D-M-Y)
Time of
Last
Product
Delivery
(local
military)
Time
between
Product
Delivery
and Data
Collection
for Test
(h-min)
Time of
Last
Dispensing
(local
military)
Time
between
Last
Dispensing
and Start of
Data
Collection
for Test
(h-min)
Time of
Temperature
Measurements
(local military)
TTB
(°F)
Ti
(°F)
T2
(°F)
T3
(°F)
TG
(°F)
TTB - TG
(°F)
Temperature
Test Matrix
Category
(Table 5.1)
-------
Attachment 4
(concluded)
Data Sheet Summarizing the Product Temperature Conditions Used in the Evaluation
Pipeline Leak Detection System
Option 2
Test No.
(Based on
Temperature
Condition)
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Date
Test
Began
(D-M-Y)
Date of
Last
Product
Delivery
(D-M-Y)
Time of
Last
Product
Delivery
(local
military)
Time
between
Product
Delivery
and Data
Collection
for Test
(h-min)
Time of
Last
Dispensing
(local
military)
Time
between
Last
Dispensing
and Start of
Data
Collection
for Test
(h-min)
Time of
Temperature
Measurements
(local military)
TTB
(°F)
Ti
(°F)
T2
(°F)
T3
(°F)
TG
(°F)
TTB - TG
(°F)
Temperature
Test Matrix
Category
(Table 5.1)
-------
Attachment 5
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Options 1 and 5
Test No. 1
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Date
Test
Began
(D-M-Y)
Induced
Leak
Rate
(gal/h)
Time between End
of Circulation and
Start of Data
Collection for Test
(h-min)
Time Data
Collection
Began
(local
military)
Time Data
Collection
Ended
(local
military)
Measured
Test Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
(continued)
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Options 1 and 5
Test No. 2
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Date
Test
Began
(D-M-Y)
Induced
Leak Rate
(gal/h)
Time between End
of Circulation and
Start of Data
Collection for Test
(h-min)
Time Data
Collection
Began
(local
military)
Time Data
Collection
Ended
(local
military)
Measured
Test
Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
(continued)
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Options 1 and 5
Test No. 3
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Date
Test
Began
(D-M-Y)
Induced
Leak Rate
(gal/h)
Time between End
of Circulation and
Start of Data
Collection for Test
(h-min)
Time Data
Collection
Began
(local
military)
Time Data
Collection
Ended
(local
military)
Measured
Test Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
(concluded)
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Options 1 and 5
Test No. 4
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Date
Test
Began
(D-M-Y)
Induced
Leak Rate
(gal/h)
Time between End
of Circulation and
Start of Data
Collection for Test
(h-min)
Time Data
Collection
Began
(local
military)
Time Data
Collection
Ended
(local
military)
Measured
Test Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Option 2
Test No. 1
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Date
Test
Began
(D-M-Y)
Induced
Leak
Rate
(gal/h)
Time between
Product
Delivery and
Data
Collection for
Test
(h-min)
Time between
Last
Dispensing
and Start of
Data Collection
for Test
(h-min)
Time Data
Collection
Began
(local
military)
Time Data
Collection
Ended
(local
military)
Measured
Test
Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
(continued)
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Option 2
Test No. 2
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Date
Test
Began
(D-M-Y)
Induced
Leak
Rate
(gal/h)
Time between
Product
Delivery and
Data
Collection for
Test
(h-min)
Time between
Last
Dispensing
and Start of
Data Collection
for Test
(h-min)
Time Data
Collection
Began
(local
military)
Time Data
Collection
Ended
(local
military)
Measured
Test
Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
(continued)
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Option 2
Test No. 3
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Date
Test
Began
(D-M-Y)
Induced
Leak Rate
(gal/h)
Time between
Product
Delivery
and Data
Collection for
Test
(h-min)
Time between
Last Dispensing
and Start of
Data Collection
for Test
(h-min)
Time Data
Collection
Began
(local
military)
Time Data
Collection
Ended
(local
military)
Measured
Test
Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
(concluded)
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Option 2
Test No. 4
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Date
Test
Began
(D-M-Y)
Induced
Leak Rate
(gal/h)
Time between
Product
Delivery and
Start of Data
Collection for
Test
(h-min)
Time between
Last Dispensing
and Start of Data
Collection for
Test
(h-min)
Time Data
Collection
Began
(local
military)
Time Data
Collection
Ended
(local
military)
Measured
Test
Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Options 3 and 4
Test No.
(Based on
Temperature
Condition)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Date Test
Began
(D-M-Y)
Date of
Last
Product
Delivery
(D-M-Y)
Time of
Last
Product
Delivery
(local
military)
Time between
Product
Delivery and
Start of Data
Collection for
Test
(h-min)
Time of Last
Dispensing
(local military)
Time between
Last
Dispensing
and Start of
Data
Collection for
Test
(h-min)
Time Data
Collection
Began
(local military)
Time Data
Collection
Ended
(local military)
Measured
Test Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 5
(concluded)
Data Sheet Summarizing the Test Results and the Leak Rates Used in the Evaluation
Pipeline Leak Detection System
Options 3 and 4
Test No.
(Based on
Temperature
Condition)
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Date Test
Began
(D-M-Y)
Date of
Last
Product
Delivery
(D-M-Y)
Time of
Last
Product
Delivery
(local
military)
Time between
Product
Delivery and
Start of Data
Collection for
Test
(h-min)
Time of Last
Dispensing
(local military)
Time between
Last
Dispensing
and Start of
Data
Collection for
Test
(h-min)
Time Data
Collection
Began
(local military)
Time Data
Collection
Ended
(local military)
Measured
Test Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 6
Data Sheet Summarizing the Test Results and the Trapped Vapor Tests
Pipeline Leak Detection System
Options 1 and 5
Summary of Temperature Conditions
Test No.
1
2
3
Date
Test
Began
(D-M-Y)
Nominal
Product
Temperature
before
Circulation
Was Started
(°F)
Time
Circulation
Started
(local
military)
Time
Circulation
Ended
(local
military)
Duration of
Circulation
(h-min)
Time of
Temperature
Measurements
(local military)
TTB
(°F)
(°F)
T2
(°F)
T3
(°F)
TG
(°F)
TTB-
TG
m
Temperature
Test Matrix
Category
(Table 5.1)
Summary of Leak Rates
Test No.
1
2
3
Date Test
Began
(D-M-Y)
Pipeline
Pressure
(psi)
Induced Leak
Rate
(gal/h)
Time between End of
Circulation and Start of
Data Collection for Test
(h-min)
Time Data
Collection Began
(local military)
Time Data
Collection Ended
(local military
Measured
Test
Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 6
Data Sheet Summarizing the Test Results and the Trapped Vapor
Pipeline Leak Detection System
Option 2
Summary of Temperature Conditions
Tests
Test
No.
1
2
3
Date
Test
Began
(D-M-Y)
Date of
Last
Product
Delivery
(D-M-Y)
Time of
Last
Product
Delivery
(local
military)
Tune
between
Product
Delivery and
Start of Data
Collection
for Test
(h-min)
Time of
Last
Dispensing
(local
military)
Time between
Start of Data
Collection for
Test and Last
Dispensing
(h-min)
Time of
Temperature
Measurements
(local military)
TTB
(°F)
Ti
(°F)
T2
(°F)
T3
(°F)
TG
(°F)
TTB - TG
(°F)
Temperature
Test Matrix
Category
(Table 5.1)
Summary of Leak Rates
Test No.
1
2
3
Date Test
Began
(D-M-Y)
Pipeline
Pressure
(psi)
Induced Leak
Rate
(gal/h)
Time between
Product
Delivery and
Start of Data
Collection for
Test
(h-min)
Time between
Start of Data
Collection for
Test and Last
Dispensing
(h-min)
Time Data
Collection
Began
(local military)
Time Data
Collection
Ended
(local military)
Measured
Test
Result
(gal/h)
Was
Threshold
Exceeded?
(yes or no)
-------
Attachment 7
Data Sheet Summarizing the Test Results Used to Check the Relationship
Supplied by the Manufacturer for Combining the Signal and Noise
Pipeline Leak Detection System
Options 1 and 5
First Check
Test No.
1
2
3
4
5
6
Actual Leak Rate*
(gal/h)
Measured Leak Rate
(gal/h)
Recommended leak rates for monthly monitoring tests and line tightness tests: 0.0, 0.05, 0.10, 0.20, 0.30 and
0.40 gal/h. Recommended leak rates for hourly tests: 0.0, 2.0, 2.5, 3.0, 3.5, and 4.0 gal/h.
Second Check
Test No.
A
B
C
A + B*
Actual Leak Rate*
(gal/h)
Measured Leak Rate
(gal/h)
A + B is the summation of the results of Tests A and B using the manufacturer's relationship for combining the
signal and the noise.
-------
APPENDIX C
Protocol Notification Form
I have received a copy of Standard Test Procedures for Evaluating Leak Detection Methods: Pipeline Leak
Detection Systems and would like to be placed on a mailing list in case changes or modifications are made
to this document.
Name:
Title:
Company:
Address:
(Street)
(City, State, Zip)
Telephone:
Mail this form to:
Office of Underground Storage Tanks
U.S. Environmental Protection Agency
Attention: Pipeline Evaluation Protocol
401 M Street, S. W.
Mail Stop OS-410
Washington, D.C. 20460
153
-------
154
-------
APPENDIX D
Random Selection of Leak Rates
Condition No. 1
Test No. Leak
Rate
(gal/h)
1 0.31
2 0.18
3 0.39
4 0.35
5 0.33
Condition No. 2
Test No. Leak
Rate
(gal/h)
1 0.20
2 0.21
3 0.25
4 0.49
5 0.37
Condition No. 3
Test No. Leak
Rate
(gal/h)
1 0.12
2 0.11
3 0.11
4 0.28
5 0.42
Condition No. 4
Test No. Leak
Rate
(gal/h)
1 0.46
2 0.36
3 0.23
4 0.17
5 0.15
Condition No. 5
Test No. Leak
Rate
(gal/h)
1 0.22
2 0.31
3 0.42
4 0.48
5 0.42
Condition No. 6
Test No. Leak
Rate
(gal/h)
1 0.39
2 0.23
3 0.26
4 0.43
5 0.11
Condition No. 7
Test No. Leak
Rate
(gal/h)
1 0.12
2 0.37
3 0.26
4 0.29
5 0.44
Condition No. 8
Test No. Leak
Rate
(gal/h)
1 0.28
2 0.41
3 0.49
4 0.47
5 0.24
Condition No. 9
Test No. Leak
Rate
(gal/h)
1 0.35
2 0.13
3 0.16
4 0.46
5 0.23
Condition No. 10
Test No. Leak
Rate
(gal/h)
1 0.34
2 0.14
3 0.35
4 0.34
5 0.20
Condition No. 11
Test No. Leak
Rate
(gal/h)
1 0.25
2 0.40
3 0.18
4 0.37
5 0.30
Condition No. 12
Test No. Leak
Rate
(gal/h)
1 0.45
2 0.10
3 0.31
4 0.30
5 0.42
Condition No. 13
Test No. Leak
Rate
(gal/h)
1 0.11
2 0.41
3 0.15
4 0.12
5 0.45
Condition No. 14
Test No. Leak
Rate
(gal/h)
1 0.49
2 0.15
3 0.42
4 0.49
5 0.21
Condition No. 15
Test No. Leak
Rate
(gal/h)
1 0.18
2 0.33
3 0.28
4 0.34
5 0.35
Condition No. 16
Test No. Leak
Rate
(gal/h)
1 0.26
2 0.25
3 0.21
4 0.14
5 0.45
Condition No. 17
Test No. Leak
Rate
(gal/h)
1 0.12
2 0.38
3 0.15
4 0.45
5 0.21
Condition No. 18
Test No. Leak
Rate
(gal/h)
1 0.16
2 0.33
3 0.46
4 0.49
5 0.40
Condition No. 19
Test No. Leak
Rate
(gal/h)
1 0.45
2 0.39
3 0.22
4 0.16
5 0.41
Condition No. 20
Test No. Leak
Rate
(gal/h)
1 0.21
2 0.33
3 0.30
4 0.14
5 0.17
Condition No. 21
Test No. Leak
Rate
(gal/h)
1 0.24
2 0.34
3 0.41
4 0.27
5 0.25
-------
156
-------
APPENDIX E
STATISTICS
Tliis appendix defines the statistical calculations that must be made in the protocol and
presents a simple example using only five data points to illustrate the calculations. Many of the
commercially available spreadsheets and most mathematical calculators have a function with
which to calculate the mean and standard deviation from a set of data and to fit a least-squares
line to these data. The confidence intervals can be easily calculated once the mean and standard
deviation are known. - ...•••
Mean and Standard Deviation
When a collection of data is being analyzed, it is often useful to examine the average value
of the data and the spread of the data around that average. These two data qualities are given
numerically by the mean and the standard deviation.
The mean, or the average, of a set of data is generally denoted by a bar over the data
variable, e.g., x, and is calculated as
- • , . N
" S X, - , • ' • - "
— _!=!•• _Xi + X2 + X3-F...+XN
x__= _ _,
where N is the number of data samples and x, is the i* data sample. £ is the symbol used to
represent the summation. . ,
The standard deviation, denoted by s, measures the spread around the mean and is ,
calculated by
s =
This equation is sometimes seen in an alternate form as
s =
where (x)2 is the square of the mean of the data and x2 is the mean of the squared data. An
example of these calculations is given in Table E.l. (Sometimes the standard deviation is
calculated with N instead of N - 1 in the denominator.)
151
-------
Table E.l. Example of Mean and Standard Deviation Calculations
1 83
2 90
3 94
4 86 ;
N«S m
Sum 435
Mean 87
16
9
49
1
25 /
100
6,889
8,100
8,836
7,396
6.724
37,945
7,589
Standard Deviation ' 5 = " 5.0 or
= "\/5 • ^if7'* = 5-0
Confidence Intervals on the Mean and Standard Deviation
The confidence interval on a quantity is the range of values which are not statistically
different from a specific value of the quantity. For example, if the confidence interval' on a mean
of 2.0 is from 1.7 to 2.5, a measured mean within the range of 1.7 to 2.5 is not statistically
different from a mean of 2.0. The confidence intervals on the mean and on the standard
deviation are calculated with the t distribution land the %2 distribution, respectively.
To calculate the 95% confidence interval on a mean, x, of N samples, we first use a
t-distribution table (found in any basic statistics book) to determine the value of t for a = 0.05
and for degrees of freedom equal to N - 1. If the standard deviation of these N samples is s, the
confidence interval is given by
• s-t
. X±-f=.
•VN
For N = 5, the value of the t-statistic for a one-tailed test is 2.78. The lower and upper
confidence intervals on the mean for the data shown in Table E.I are 80.784 and 93.216,
respectively.
To calculate the 95% confidence interval on the standard deviation, we first use a
%2-distribution table to determine the values of %2 for a = 0.05 and for 1 - a = 0.95, both for N - 1
degrees of freedom. The lower limit of the confidence interval is then given by
and the upper limit is given by
152
-------
Values for the %2-distribution can be obtained in the appendices of most statistics
textbooks. The lower and upper confidence intervals on the standard deviation for the data
shown in Table E.I are 3.627 and 13.259, where x20tQS = 9.500 and %20.95 = 0.711 for 4 degrees of
freedom.
Linear Regression Analysis: Least-squares Fit
In studying the relationship between two measured quantities, it is desirable to derive from
experimental data an equation that best expresses this relationship. For cases in which the data
seem to be linearly related, a best fit to the data is obtained by using the linear regression method
of least squares.
Let the 1th value of the independent data variable be x-s and the corresponding dependent
data variable be yt. Then, the linear relationship between x and y is given by
y = mx + b . .
where
_ __.
yx2
x2
_
x1
- (x)2
/ N
\i "~ *
- (x)2
•\
ft x
J
and N is the number of data pairs. (For an explanation of x, see the section at the beginning of,
this appendix entitled Mean and Standard Deviation).
Two different quantities are used as a measure of the accuracy of the linear fit. The first is
the variance* along the regression line given by
N N. N-
Zyf - b Zy, -
2 _ i=li=I
* ~
N - 2
The second measure of the accuracy is the variance of the slope given by
NS2
A least-squares line was fit to the data in Table E.1; the results show that m = -0.600, b =
88.800, s = 5.699, and sm = 1.793. An x-y plot of the data shown in Table E.I will show that the
data are not modeled well by a line.
* "The variance is simply the standard deviation squared.
153
-------
APPENDIX E
STATISTICS
This appendix defines the statistical calculations that must be made in the protocol and presents a simple
example using only five data points to illustrate the calculations. Many of the commercially available
spreadsheets and most mathematical calculators have a function with which to calculate the mean and
standard deviation from a set of data and to fit a least-squares line to these data. The confidence intervals
can be easily calculated once the mean and standard deviation are known.
Mean and Standard Deviation
When a collection of data is being analyzed, it is often useful to examine the average value of the data and
the spread of the data around that average. These two data qualities are given numerically by the mean
and the standard deviation.
The mean, or the average, of a set of data is generally denoted by a bar over the data variable, e.g., , and
is calculated as
where N is the number of data samples and Xj is the ith sample. is the symbol used to represent the
summation.
The standard deviation, denoted by , measures the spread around the mean and is calculated by
This equation is sometimes seen in an alternate form as
where is the square of the mean of the data and is the mean of the squared data. An example of
these calculations is given in Table E.1. (Sometimes the standard deviation is calculated with N instead of
N -1 in the denominator.)
157
-------
Table E.1. Example of Mean and Standard Deviation Calculations
i
1
2
3
4
N = 5
Sum
Mean
83
90
94
86
82
435
87
16
9
49
1
25
100
6,889
8,100
8,836
7,396
6.724
37,945
7,589
Standard Deviation
— or
Confidence Intervals on the Mean and Standard Deviation
The confidence interval on a quantity is the range of values which are not statistically different from a
specific value of the quantity. For example, if the confidence interval on a mean of 2.0 is from 1.7 to 2.5, a
measured mean within the range of 1.7 to 2.5 is not statistically different from a mean of 2.0. The
confidence intervals on the mean and on the standard deviation are calculated with the t distribution and the
X2 distribution, respectively.
To calculate the 95% confidence interval on a mean, , of N samples, we first use a t-distribution table
(found in any basic statistics book) to determine the value of t for a = 0.05 and for degrees of freedom equal
to N - 1. If the standard deviation of these N samples is , the confidence interval is given by
For N = 5, the value of the t-statistic for a one-tailed test is 2.78. The lower and upper confidence intervals
on the mean for the data shown in Table E.1 are 80.784 and 93.216, respectively.
To calculate the 95% confidence interval on the standard deviation, we first use a x2-distribution table to
determine the values of x2 for a = 0.05 and for 1 - a = 0.95, both for N -1 degrees of freedom. The lower
limit of the confidence interval is then given by
and the upper limit is given by
158
-------
Values for the x2-distribution can be obtained in the appendices of most statistics textbooks. The lower and
upper confidence intervals on the standard deviation for the data shown in Table E.1 are 3.627 and 13.259,
where = 9.500 and = 0.711 for 4 degrees of freedom.
Linear Regression Analysis: Least-squares Fit
In studying the relationship between two measured quantities, it is desirable to derive from experimental
data an equation that best expresses this relationship. For cases in which the data seem to be linearly
related, a best fit to the data is obtained by using the linear regression method of least squares.
Let the ith value of the independent data variable be Xj and the corresponding dependent data variable be yi.
Then, the linear relationship between x and y is given by
where
and N is the number of data pairs. (For an explanation of , see the section at the beginning of this
appendix entitled Mean and Standard Deviation).
Two different quantities are used as a measure of the accuracy of the linear fit. The first is the variance*
along the regression line given by
The second measure of the accuracy is the variance of the slope given by
A least-squares line was fit to the data in Table E.1; the results show that m = -0.600, b = 88.800, s = 5.699,
and sm = 1.793. An x-y plot of the data shown in Table E. 1 will show that the data are not modeled well by a
line.
The variance is simply the standard deviation squared.
159
------- |